NORWEGIAN JOURNAL OF GEOLOGY Diagenesis of the Minkinfjellet Formation, Central Spitsbergen 319
Eliassen, A. & Talbot, M. R.: Diagenesis of the mid-Carboniferous Minkinfjellet Formation, central Spitsbergen Svalbard.
Norwegian Journal of Geology , Vol. 83, pp. 319-331. Tromsø 2003. ISSN 029-196X.
The Moscovian Minkinfjellet Formation, deposited in the NNE-SSW striking Billefjorden Trough in central Spitsbergen, is a carbonate-dominated succession with interbedded sandstones and evaporites. The sediments accumulated in a narrow, shallow marine basin where tidal currents and high frequency sea-level changes controlled sediment distribution and depositional environments. After deposition the succession was influenced by a series of diagenetic effects, and the carbonate rocks, in particular, were altered by several different processes. Three main generations of calcite cement have been recognised from both petrographic and stable isotope studies. These are: (i) early marine fibrous cement, possibly originally aragonitic (ii) drusy spar cement formed during flushing by meteoric water, and (iii) coarse, crystalline pore-filling cement precipitated during deeper burial of the rocks.
δ
18 O values from these cements become successively isotopically lighter, and show typical values for each of these diagenetic environments. Complete or partial dolomitization is common and widespread, presumably caused by flushing of hypersaline Mg-rich brines. The abundance of interbedded gypsum in this succession supports a seepage-reflux model of dolomitization. Chert nodules and chert layers are most common in the carbonates, both limestones and dolomites. Three mechanisms of silicification are recognised: (i) Replacement of gypsum nodules by silica. These show growth of quartz crystals toward the nodule centre, where some gypsum may still remain, (ii) Concretionary growth of finely crystalline glassy chert nucleated around organic matter, which has induced the precipitation of silica, (iii) Intergranular megaquartz formation in ooidal or peloidal limestones. In these ooid grainstones calcite spar was effectively replaced by megaquartz due to infiltration of silica-rich solutions through crystal defects or along crystal contacts. The distribution of chert in the Minkinfjellet Formation rocks was controlled by the initial accumulation of siliceous distribution of silica sponge spicules.
Quartz cement has developed in all investigated sandstones in the form of quartz overgrowth around grains. Many sandstones show pore-filling calcite cement engulfing the quartz cement, effectively reducing the original porosity.
Arild Eliassen 1 & Michael R. Talbot, Department of Earth Sciences, University of Bergen, N-5007 Bergen, Norway.
N-4035 Stavanger, Norway.
1 Present address: Statoil ASA,
The Moscovian Minkinfjellet Formation is a carbonatedominated succession consisting of interbedded sandstones, evaporites and minor shales situated in the central part of Spitsbergen, Svalbard (Figs. 1 and 2). Solution collapse breccias are common and widespread within the formation and were formed during dissolution of interbedded evaporites (Eliassen 2002; Lønøy 1995;
McWhae 1953). The aim of this paper is to give an overview of the diagenetic processes that have affected the carbonates and siliciclastics of the Minkinfjellet
Formation. The gypsum karst history will not be described in detail in this study, but will be presented elsewhere (Eliassen 2002). The diagenesis affecting the unkarstified parts of the succession involves several phases of calcite cementation, partial or total dolomitization, aggrading neomorphism, and several mechanisms of silicification and silica cementation. These diagenetic effects have in general greatly altered the
Minkinfjellet Formation rocks. In addition to describing the sedimentary facies and depositional history of the formation (Eliassen 2002) it is important to focus on the diagenetic history in order to understand completely the evolution of a complex succession like the
Minkinfjellet Formation.
Primary carbonate facies of the Minkinfjellet Formation are dominated by shallow-water, open-marine wackestones/packstones and ooid and skeletal grainstones, with interbedded restricted marine or lagoonal mudstones and tidal flat algal laminites (Eliassen 2002; Lønøy
1995). The basin-fill siliciclastics are dominated by tidal and alluvial sandstones, in contrast to the thick wedges of sandstones and conglomerates of the underlying
Ebbadalen Formation (Johannessen & Steel 1992).Tidal
currents and rapid sea-level changes in the narrow Billefjorden Trough were the dominant controlling factors on the distribution and evolution of the sediments of the Minkinfjellet Formation (Eliassen 2002)
The Minkinfjellet Formation was deposited in the Billefjorden Trough, a NNE-SSW oriented half-graben pro-

320 NORWEGIAN JOURNAL OF GEOLOGY A. Eliassen & M. R. Talbot
Fig. 1. Map showing the area of outcropping Minkinfjellet Formation, position of logged profiles and key localities.

NORWEGIAN JOURNAL OF GEOLOGY Diagenesis of the Minkinfjellet Formation, Central Spitsbergen 321
Fig. 3. Cross-section of the Billefjorden Trough and adjacent structural highs; The Minkinfjellet Formation represents the late syn-rift sediments in the trough. Modified from Johannesen and Steel
(1992).
Fig. 2. Scheme showing the age and stratigraphic position of the
Minkinfjellet Formation. After Dallman et al. (1999).
blocks and development of a wider, open-marine carbonate platform covering large parts of Spitsbergen
(Cutbill & Challinor 1965; Pickard, et al. 1996).
duced by oblique-slip faulting during Bashkirian to
Gzhelian times (Gjelberg & Steel 1981, Steel & Worsley
1984)(Fig. 3). The Minkinfjellet Formation is of Moscovian age and thus comprises some of the uppermost syn-rift sediments of the Billefjorden Trough (McCann
& Dallmann 1996; Pickard et al., 1996; Samuelsberg &
Pickard 1999). The underlying Ebbadalen Formation consists of a series of clastic wedges shed from the
Nordfjorden Block, a structural high which acted as a major source area to the west, and evaporites (gypsum and anhydrite) eastwards in the basin (Fig. 3). During deposition of the overlying Minkinfjellet Formation a transition from mixed siliciclastic, evaporitic and carbonate deposition into pure carbonate deposition occurred. The lower unit of the Minkinfjellet Formation, the Carronelva Member, consists of a heterolithic succession containing interbedded sandstones, carbonates and minor evaporites. The Carronelva Member was deposited in a mixed siliciclastic and carbonate basin with occasional evaporative conditions, where increased carbonate dominance was controlled by progressive reduction in the relief of the source areas to the east and west (Eliassen 2002).
The transition to the overlying Terrierfjellet member was marked by a major sea-level fall, with ensuing erosion and scouring in the northern basin and a transition to thick lowstand evaporites in the southern basin
(Eliassen 2002). After the subsequent transgression a pure carbonate platform developed, with deposition of mainly open-marine limestones, but still limited by the margins of the Billefjorden Trough. The transition to the overlying Wordiekammen Formation represents the final drowning of the Nordfjorden and Ny-Friesland
Logging and sampling of the Minkinfjellet Formation were carried out during July and August 1999 and July and August 2000. Thin section studies were made with plane-polarised transmitted light with a Nikon type
119 microscope. Cement and whole-rock sampling for carbon and oxygen stable isotopic analysis was carried out with a jeweller’s drill. All stable isotopic analyses were performed using conventional methods at the
GMS Laboratory, Department of Geology, UIB, on a
Finnegan Mat 251 mass spectrometer. Results are expressed as per mil (‰) deviation from the Pee Dee
Belemnite (PDB).
Several diagenetic processes have affected the carbonate rocks of the Minkinfjellet Formation. These include different stages of calcite cementation, aggrading neomorphism, dolomitization and silicification.
Calcite cements
Description
Three calcite cement types has been recognised in the
Minkinfjellet Formation limestones: (i) thin rims of radial fibrous calcite are common around skeletal grains (Fig. 4A), (ii) Equant spar showing a drusy mosaic fabric is common in many oolitic and skeletal grainstones and wackestones. This spar is located at or near grain contacts, or as cement filling of the primary

322 NORWEGIAN JOURNAL OF GEOLOGY
Fig. 4. A) Photomicrograph of limestone showing early marine fibrous cement and equant late spar surrounding and filling a fossil fragment. B) Photomicrograph of ooid limestone containing meteoric cement. This type of cement is developed close to exposed surfaces.
porosity (Fig. 4B). The cement type is found at or close to detected exposure surfaces (microkarst structures,
Microcodium , desiccation cracks) and (iii) coarse blocky spar filling larger primary or secondary voids.
Oxygen and carbon isotope analyses have been performed on samples containing the different calcite cements described above. It was difficult to isolate pure cement samples of the radial fibrous cement and the drusy mosaic spar, thus these results are based on whole rock samples of limestones containing these cements.
Significant compositional differences are, however, apparent. Whole rock samples of marine wackestone and packstones containing rims of fibrous cements with little or no evidence of other generations of calcite cements, show
δ 18 O values varying from –0.46 to –6.50
‰ and
δ 13 C values ranging from +0.12 to +3.17
‰(Fig. 5). Samples containing drusy mosaic cement
(this cement is commonly porefilling and comprises
15-30% of total rock volume) show a narrow range in
δ 18 O values, ranging from –6.83 to –7.76 ‰ and
δ 13 C ranging from +0.47 to +2.85 ‰. Two samples of coarse blocky spar were analysed, and yielded
δ 18 O values of
–9.75 and –8.73 ‰ and
δ 13 C values of +1.05 and –0.11
‰ respectively.
A. Eliassen & M. R. Talbot
Interpretation
Radial fibrous cement forming rims around skeletal grains commonly develop in a phreatic marine environment at the water-sediment interface (Tucker &
Wright 1990). This cement is therefore interpreted to represent an early cement generation precipitated shortly after deposition.
The whole rock samples of marine limestones containing only this type of cement show a rather large range in stable isotope values (Fig. 5). All values are, however, within the range of typical values of ancient marine limestones. Oxygen isotope values are more negative compared to a pure marine signature, which would approach zero. This is considered to reflect several possible factors. The oxygen isotopic composition of oceanic waters has changed through time, becoming isotopically heavier towards the present day (Veizer et al.
1997; Popp, et al. 1986; Veizer et al. 1986; Fallick &
Hamilton 1989). This may explain the negative values of some samples. The whole rock samples will yield a weighted mean value of all components of the rock, which also may include isotopic effects produced by flushing of meteoric waters and during burial. The
δ 18 O value may thus represent a mixture of the isotopic imprint from several diagenetic stages (although these are not necessarily detected in thin-sections), and as a result the values are depleted with respect to
δ 18 O
(Given & Lohmann, 1985). Together these whole-rock samples define a mixing-line between the pure marine end-member of the sediment and the effects of later diagenetic phases. The positive carbon isotopic composition of these samples is also typical for Carboniferous marine limestones which in general show a "heavy" signal, possibly due to enrichment of 13 C in ocean waters caused by increased burial of isotopically light carbon at that time (thick Carboniferous coal deposits (Popp et al. 1986; Veizer et al. 1986).
The drusy mosaic spar is interpreted to represent meteoric cement. The fact that it is only found close to exposed surfaces supports this interpretation. The isotopic data from samples taken close to exposed surfaces are consistent with the inferred meteoric origin of the drusy spar seen in these samples. Oxygen isotope values are depleted, showing a narrow range (averaging –7.36
‰ and maximum and minimum values are separated by less than 1 ‰), while carbon isotopes show a wider range of values. A marine origin for this isotopic signature is not likely. Although the
δ 18 O of Carboniferous seawater was probably lower than modern seawater (1-
2 ‰) (Veizer et al. 1997), it would not have been possible to precipitate such isotopically depleted calcite cement from Carboniferous seawater at normal surface temperatures. The isotopic signature is, however, typical for meteoric calcite cements, defining a "meteoric calcite line" (Allan & Matthews 1982; Lohmann, 1988).

NORWEGIAN JOURNAL OF GEOLOGY Diagenesis of the Minkinfjellet Formation, Central Spitsbergen 323
Fig. 5. Cross-plot of oxygen and carbon stable isotopes of cements and whole rock samples of the Minkinfjellet Formation carbonates. Meteoric cements show a typical meteoric calcite-line confirming the origin of drusy spars. Unaltered limestones, neomorphic limestones and dolomites show values representing a mixing-line between typical Carboniferous marine values and later diagenetic states. Late spars show depleted oxygen and carbon isotopic values, which are typical for cements precipitated at elevated temperatures. See text for more detailed interpretation and discussion.
Despite a general lowering of carbon isotope values compared to the marine values, the carbon isotopic composition of the blocky cement is still rather heavy
(close to that expected for calcite precipitated from
Carboniferous seawater (Veizer et al. 1986). This is most likely due to the limited contribution of isotopically light soil CO
2 to the deep groundwater system.
The periods of exposure responsible for the development of a meteoric phreatic lens probably did not last long enough to produce mature soil covers (the exposure surfaces are evidenced by immature karst surfaces, desiccation cracks or microcodium development (cf.
Eliassen 2002). Paleosols, root-structures or coal fragments typically developed at longer-lived exposure surfaces are absent, thus soil CO
2 was rapidly consumed during the initial phases of rock/water interaction.
Only two stable isotope measurements of coarse blocky cement have been carried out, and these are not sufficient for conclusive interpretations. However, the stableisotope composition of coarse blocky spars shows further depletion of
δ 18 O compared to the other calcite phases. These spars were probably precipitated during deeper burial of the limestones, and the negative oxygen values are considered to be the result of precipitation at elevated temperatures (Choquette & James
1986). The carbon isotopic composition of these spars shows a slight depletion in 13 C compared to marine limestones (Fig 5), which generally takes place during burial and increasing temperatures (Choquette &
James 1986). The coarse blocky spar isotope data are thus consistent with typical burial calcite cements.
Neomorphism
Description
Many limestones in the Minkinfjellet Formation have a homogeneous crystalline appearance, with sparse fossils and no sedimentary structures (Fig. 6A). These limestones may be confused with dolomites because of their crystalline appearance. Thin-section studies, however, reveal that these rocks are composed entirely of calcite microspar. Crystal sizes vary from 10 to 100
µ m. Few or no primary features are preserved, apart from scattered large crinoid or brachiopod fragments.
Oxygen and carbon stable isotopes have been measured

324 NORWEGIAN JOURNAL OF GEOLOGY A. Eliassen & M. R. Talbot growth is linked to the removal of Mg from the crystal lattice of the original lime-mud. Clay minerals, especially chlorites and montmorillonite can adsorb Mg-ions released by high Mg-calcite (Longman 1977). The presence of clays in the micrite is thus a controlling factor for the neomorphmism. Baush (1968) showed that neomorphism is restricted to limestones with less than
2 % clay. Higher clay contents thus prohibit neomorphism. Flushing of fresh meteoric waters is considered necessary for this process to initiate (Folk 1964). The recrystallization process is, however, thought to be a subsurface phenomenon, taking place long after deposition (Longman 1977) (Table 1).
The oxygen isotopic values are similar to, and show a similar mixing line to the unaltered limestones (Fig. 5), thus the same explanation for this trend can be applied to the neomorphic limestones. Carbon isotopic values show little variation and are consistent with typical values for marine limestones of this age. Neomorphism is considered to be a closed system process, thus no isotopic exchange is expected to have occurred during the recrystallization process. The isotopic composition of these rocks will be the same as that of the original rock prior to neomorphism.
Fig. 6. A) Photomicrograph of pseudosparitic limestone. Neomorphism of lime-mudstones and wackestones after deposition produces crystalline limestones. Fossil fragments may be preserved during neomorphism (arrow). B) Photomicrograph of coarse crystalline dolomite. Example of complete dolomitization showing no features of the original sediment. Dolomite neomorphism during burial is possibly responsible for the coarse crystallinity. C) Example of partial dolomitization of limestones.
in four samples of neomorphic limestones (Fig. 5) and d18O values range from –0.95 to –4.50 ‰ and
δ 13 C values from +1.56 to +2.05 ‰.
Interpretation
Development of microspar/pseudospar is commonly associated with recrystallization of fine-grained micritic limestones (Folk 1964; Longman 1977). Crystal
Dolomitization
Description
Dolomitization is the volumetrically most important diagenetic process that has occurred in the basin. Large quantities of limestone have been completely or partially dolomitized. The dolomite rocks are dark grey, crystalline and typically have a bright yellow colour on weathered surfaces. Macrofossils such as crinoids, brachiopods and gastropods are commonly preserved, but microfauna or matrix components are rarely or never preserved. Dolomite crystals vary in size from ca 20 to
500
µ m (Fig. 6B). Both idiotopic and xenotopic dolomite are common and many limestones show varying degrees of partial dolomitization. (Fig. 6C).
Stable isotope data of dolomites show
δ 18 O values varying from –2.90 to –6.89 ‰ and
δ 13 C values ranging from +2.12 to +3.85 ‰ (Fig. 5).
Interpretation
Dolomitization of limestones requires an elevated porewater Mg/Ca. One of the most common and accepted models for dolomitization is seepage-reflux, where high Mg/Ca ratios in pore-waters are generated by evaporation beneath tidal flats and sabhkas. Because of the density contrast with marine porewaters the saline brines seep downwards through the underlying deposits (Adams & Rhodes 1960). This model can be applied to the Minkinfjellet Formation dolomites (see discussion), because evaporative environments and

NORWEGIAN JOURNAL OF GEOLOGY Diagenesis of the Minkinfjellet Formation, Central Spitsbergen 325 isotopes, and tends to stay unchanged during crystal growth.
Considering these facts it is difficult, based on stable isotopes, to be conclusive regarding the composition and origin of the dolomitizing fluids, or the original isotopic composition of the dolomitized rock. The lowest dolomite δ 18 O values, however, may be isotopically depleted due to precipitation in a mixing zone environment (Muchez & Viaene 1994), i.e. the zone where mixing of fluids from the different sources occurs.
Dolomitization of the Minkinfjellet Formation strata is thus interpreted to have taken place at moderate burial depth during or shortly after evaporative conditions developed in the basin (Table 1).
Fig. 7. A) Example of a dolomitization front typically seen in the
Minkinfjellet Formation carbonates. The yellow (upper) part is dolomite, the lower an oolitic limestone. The dolomitization process is controlled by flushing by Mg-rich brines produced during deposition of adjacent gypsum beds, the primary permeability of the rocks being dolomitized. B) Photomicrograph of the same contact as shown above.
deposition of gypsum in sabhka environments were among the most common and widespread sedimentary features during the accumulation of the Minkinfjellet succession (Eliassen 2002). Dolomitization was incomplete, however, and several dolomitization fronts are observed within limestones (Fig. 7). Thin-section studies of these contacts reveal that dolomitization postdated marine and meteoric calcite cementation (Fig. 6A and B). During burial, recrystallization may create xenotopic fabrics or the growth of coarse crystalline dolomites, as seen in some beds (Fig. 6B)(Gregg &
Sibley 1984; Kupecz et al. 1998).
Elevated temperatures and dilution by meteoric waters during dolomitization will favour incorporation of 16 O into the solid phase (Mattes & Mountjoy 1980). Many ancient dolomites, however, have bulk δ 18 O values lower than those of modern dolomites (Land 1980).
There does seem to be a correlation between increasing crystal size and decreasing δ 18 O values in ancient dolomites (Land et al. 1975; Morrow 1982). Carbon isotopic composition does not show the same relation as oxygen
Silicification
Description
Chert nodules and discontinuous chert bands are the most striking macroscopic diagenetic features present in the Minkinfjellet Formation carbonates. Nodules, which are the more common, are abundant in both fine-grained limestones and dolomites, occurring throughout the formation in every part of the basin (Fig.
8A). The silica nodules are black, white or yellow, and typically occur along bedding planes and in restricted horizons. In many limestones, however, the nodules seem more randomly spaced. Nodules typically range from 5-15 cm in diameter, but smaller and larger nodules are also present. Discontinuous chert-bands have thicknesses in the same range, but may be thicker (up to 30 cm). The lateral continuity of these bands is generally less than ten metres. Most nodules are ellipsoidal in shape, with their longest axes parallel to the bedding
(subhorizontal), but the bigger ones commonly have a more irregular shape. Most nodules are finely crystalline, almost glassy and purely siliceous, but some have a core of gypsum or crystalline calcite, or they are hollow.
The latter type shows quartz crystals facing the centre of the nodule with increasing crystal size (up to centimetre scale) towards the centre. This indicates that the growth of the nodule started at the edges and propagated towards the centre.
Interpretation
Several models have been proposed for the formation of chert nodules, and the issue of timing and formation processes is much debated in the literature. Silicification of former gypsum nodules is an important mechanism forming chert nodules (Siedlecka 1972; Tucker
1976; Milliken 1979). This process is presumably responsible for the formation of many of the nodules described here. Many chert nodules contain remnants of anhydrite or gypsum in the centre of the nodule, or

326 NORWEGIAN JOURNAL OF GEOLOGY as inclusions seen in thin-section. The evaporite nodules were replaced by growth from the outer edge toward the centre, which is the preferred mechanism of gypsum replacement (Milliken 1979). The presence of nonsilicified gypsum in many limestones and dolomites, often in association with chert nodules, points in favour of this process being the replacement mechanism.
Many other chert nodules do not show the above-described features that indicate silicification of former gypsum nodules, hence other models of silicification must also be considered. Knauth (1979) proposed that growth of chert nodules preferably takes place in the mixing zone between meteoric and marine porewater.
A problem with this model is that dissolved calcite not replaced by silica is rarely seen adjacent to chert nodules. To explain this, a volumetrically perfect balance between the dissolution and precipitation process is necessary. Maliva & Siever (1989) propose that undersaturation of calcite is not necessary for the formation of chert nodules such that calcite dissolution occurs at grain contacts and not by dissolution by undersaturated pore-waters, and that chert nodules preferably
Fig 8. A) Closely spaced subrounded or elongated chert nodules are very abundant within Minkinfjellet Formation limestones and dolomites. The picture shows a grey wackestone containing black and yellowish chert nodules. Hammer shaft is 30 cm long. B) Photomicrograph of peloidal limestone containing intergranular megaquartz.
Megaquartz has effectively replaced former calcite spar which originally cemented the rock.
A. Eliassen & M. R. Talbot nucleate around, and grow from fragments of organic matter. This again may explain the widely observed silicification of skeletal fragments and trace fossils (a common feature in the Minkinfjellet Formation). The latter model does not require a preferred diagenetic environment as long as a source of silica is present. Thus the process responsible for chert nodule formation may take place at any stage in the limestone’s burial history.
However, chert nodules or fragments of nodules are common clasts within the breccias of the Minkinfjellet
Formation, which were formed in the shallow subsurface shortly after deposition (less than 500 meters burial; Eliassen 2002) This implies that the formation of chert nodules occurred, at least partially, as an early diagenetic phase in a shallow burial realm (Table 1).
A few beds of ooid/peloid grainstones within the Carronelva member on the northeastern part of Campbellryggen are cemented by silica in the form of megaquartz crystals (Fig. 8B). The ooids and peloids themselves show no signs of silicification. Formation of megaquartz cement is generally associated with evaporitic successions which implies a causal relationship in the way that evaporites are effectively replaced by silica after being dissolved (Pittman & Folk 1971; Tucker
1976; Milliken 1979; Chafetz & Zhang 1998). The pore spaces filled with megaquartz in the ooid grainstones of the Minkinfjellet Formation show no evidence of being formerly occupied by evaporites. Thus an evaporite replacement mechanism is unlikely for these cherty limestones. Bustillo et al. (1998) described megaquartz cemented oolites from southern Spain showing very similar textures. They concluded that the megaquartz had replaced original calcite-spar due to silica-rich solution moving by diffusion preferentially along the contacts between calcite crystals or through crystal defects. Ooids and allochems in these limestones were impermeable to percolating silica solutions, thus the coarse crystalline spar was preferentially replaced in the chertification process. Sutured contacts showing evidence of dissolution between ooids and peloids are common, indicating that the ooid grainstones were compacted prior to cementation. The original calcite cement was thus formed after compaction and burial.
The subsequent replacement by megaquartz is considered to be a late stage diagenetic process, taking place after deep burial (Table 1).
Implications on silicification
Three mechanisms of silicification are recognised in the
Minkinfjellet Formation carbonates: (i) Replacement of evaporite nodules, (ii) Concretionary growth of silica nucleated around organic matter, and (iii) Replacement of calcite spar by megaquartz. Any silicification model requires an excess of fluids supersaturated with respect to a silica mineral phase (opal A, opal CT, quartz). The origin of silica brines is commonly linked to the primary composition of the silicified beds, or nearby

NORWEGIAN JOURNAL OF GEOLOGY Diagenesis of the Minkinfjellet Formation, Central Spitsbergen 327 dance of primary biogenic silica in the chert nodule horizons (see Fig. 9A). Considering this fact, the cherty horizons in these carbonates imply major blooms of silica-producing organisms during deposition. Thus the abundance and distribution of chert have a direct relationship to changes in the primary depositional environment.
Dedolomitization
Description
The clasts of the collapse breccias of the Minkinfjellet
Formation are rarely composed of dolomite. This is due to effective dedolomitization controlled by flushing of
Ca 2+ into the porewater during gypsum dissolution
(Blount & Moore 1969; Moore et al. 1969; Karakitsios
& Pomoni-Papaioannou 1998). The process of dedolomitization will be described in more detail in papers concerning the breccias of the Minkinfjellet Formation
(Eliassen 2002). The dedolomitization is only observed within the Minkinfjellet Formation breccias, but it is likely that this process has also affected some of the unbrecciated dolomites in the formation, presumably dolomite beds situated close to the breccia beds.
Fig. 9. A) Siliceous sponge spicules are abundant and closely spaced in many limestones of the Minkinfjellet Formation. Arrows show examples of spicules in cross-section and longitudinal section. (Crossed polars). B) Another example of preserved siliceous spicules in cross-section (arrows). Sponge spicules are considered to be the silica source for chertification of the Minkinfjellet Formation carbonates.
deposits. The source of silica in ancient carbonate successions has been much debated. However, most commonly the silica is thought to be sourced from dissolved siliceous skeletal fragments, such as sponge spicules and radiolarians (Gimenés-Montsant et al. 1999; Knauth
1979; Kwiatkowski 1981; Makhnach & Gulis 1993;
Maliva & Siever 1989; Milliken 1979; Pittman & Folk
1971). This is also likely to have been the case for the
Minkinfjellet Formation where silica sponge spicules have been identified in thin-sections (Figs. 9A and B).
Maknach & Gulis (1993) commented on an important relationship in chert-bearing carbonates, which is that the abundance of chert nodules in many limestones and dolomites is very high, comprising a significant portion of the total rock. They conclude that as much as 10% of the initial sediment must have been composed of siliceous skeletal fragments to have been able to produce such large amounts of chert in the rock. The spacing of chert nodules in the Minkinfjellet carbonates is generally very close, typically less than a half metre separating nodules of 10 cm mean size (ca 20% by volume) (Fig. 8A), which implies a similar abun-
Description
Most of the sandstones present in the Minkinfjellet
Formation contain calcite cement (Fig. 10A), which fills intergranular pore space and is a typical drusy calcite spar. Sutured grain contacts with evidence of grain contact dissolution are very abundant between quartz grains. Syntaxial overgrowth of silica cement is common, including samples containing calcite spar (Fig.
10B). Many sandstones are free of carbonate cement and show high intergranular porosities (above 20%)
(Fig. 10B).
Interpretation
Early cementation by calcite spar will prohibit dissolution and precipitation of quartz cement (Tucker 1991), and hence it is thought to have precipitated after the quartz cement. Carbonate cementation is therefore considered to be late, formed after deep burial. No isotopic data are available for this carbonate cement, and therefore this interpretation is based solely on petrographic observations. A source of carbonate is not difficult to find in the Minkinfjellet Formation where limestones are the dominant lithology. Dissolution and reprecipitation during burial of adjacent carbonate deposits and of skeletal fragments within the sands probably provided the principal source of calcite cements in the Minkinfjellet Formation sandstones.

328 NORWEGIAN JOURNAL OF GEOLOGY
Fig. 10. A) Sandstone containing calcite cement effectively reducing porosity. B) Sandstone showing quartz cementation (crystal overgrowth) and sutured grain contacts.
Sequential evolution of diagenetic events in the Minkinfjellet Formation.
There are no obvious petrographic similarities between the carbonate cement in the sandstones and the latest spars in the carbonate rocks. The late spars of the carbonate rocks are coarser and show better developed equant crystallinity. The timing and depth of burial of these cements are, however, thought to be approximately simultaneous (Table 1).
A. Eliassen & M. R. Talbot
There is, in general, a close relationship between depositional environment and the subsequent diagenesis of a carbonate succession. The diagenetic effects taking place during the burial of a sediment are in many ways controlled by the primary constituents of the sediment itself, or by the overlying or underlying deposits. In the case of Minkinfjellet Formation, dolomitization and chertification are two examples of this relationship.
The dolomitization history of the Minkinfjellet Formation can be related to the primary deposits. Percolating porewaters with elevated Mg/Ca ratios were responsible for the extensive dolomitization of the carbonates.
These brines were produced while evaporative conditions dominated at the sediment surface, resulting in gypsum and anhydrite deposition. The degree of dolomitization increases upward in the succession, in parallel with the increase in evaporite deposits. Due to intense, later karstification of these evaporite beds this relationship is not always obvious, because many of the gypsum beds have now been removed by dissolution.
Their former distribution in the basin is instead represented by extensive collapse (Eliassen 2002). The production of Mg-rich brines responsible for dolomitization depends upon in which setting the evaporites are deposited in. A pure sabhka origin in a supratidal setting where displasive gypsum grows within the sediment in conditions of hypersaline porewaters, is often associated with dolomitization of the surrounding carbonates (Maliva 1987). However, under these conditions the zone of dolomitization is generally localised close to the gypsum deposit itself, and represented by dolomitic crusts in sabhka or lagoonal sediments (Tucker & Wright 1990). Dolomitization of a thick carbonate succession requires rather large amounts of Mgrich brines, thus pure sabhka conditions which occasionally develop at the basin margins may not produce sufficient amounts of dolomitizing fluids. Kendall &
Harwood (1996) argue that thicker beds of gypsum and anhydrite (despite being nodular) most likely originated in a subaqueous environment, such as lagoons and perennial salinas, with prolonged restricted circulation and hypersaline seawater. The thicker gypsum beds present in the Minkinfjellet Formation are thought to have such a subaqueous origin. In this setting significant amounts of water enriched with respect to Mg were produced in the basin during deposition of the evaporites. The latter condition would have been capable of dolomitizing thick carbonate successions like those in the Minkinfjellet Formation.
Dolomitization is also a common effect in settings where marine and meteoric waters mix (Land 1973;
Choquette & Steinen 1980; Muchez & Viaene 1994).
Mixing-zone conditions may well have occurred during deposition of the Minkinfjellet Formation. Exposure

NORWEGIAN JOURNAL OF GEOLOGY surfaces and meteoric cements in many limestones provide evidence of meteoric porewater penetrating the succession during sea-level lowstands. Mixing zones are there a natural consequence of such conditions. Widespread dissolution of the interbedded gypsum, with subsequent gravitational collapse forming carbonate breccias, imply extensive flushing by dilute waters and the development of a mixing zone over large areas of the basin subsurface (Eliassen 2002). Dolomitization by mixed marine and meteoric porewaters cannot therefore be ruled out in the Minkinfjellet Formation, but an important observation argues against this process having been of major significance. The collapse breccias found in the formation are more or less completely dedolomitized. This effect leads to two conclusions i) dolomitization took place prior to the development of fresh-water lenses in the formation (implying that dolomitization took place during deposition of the evaporites by hypersaline brines), and ii) flushing by dilute waters will in this case cause dedolomitization, not dolomitization, because of the high abundance of gypsum/ anhydrite in the formation.
Extensive dolomitization during deeper burial has been documented in several studies (Mattes & Mountjoy
1980; Morrow 1982). As for the mixing zone model, this mechanism is applied to successions lacking evidence of evaporites or access of hypersaline evaporative fluids. The Mg-rich brines responsible for dolomitization in a burial realm are thought to be sourced from formational waters expelled from shales during compaction, where Mg-ions are derived from clay-mineral transformations. The Minkinfjellet and surrounding formations do not contain significant amounts of shale: only thin (typically 10-50 cm thick) scattered beds are found within the succession (Eliassen 2002).
Thus dolomitization during deeper burial is not considered as an important factor in the dolomitization history of the Minkinfjellet Formation. However, dolomite recrystallization forming coarse crystalline, sucrosic dolomites (as described earlier) most likely took place in many places during later burial.
Diagenesis of the Minkinfjellet Formation, Central Spitsbergen 329 pretations. Isotopic values for marine cements reflect the composition of Late Carboniferous marine waters. Meteoric cements show a negatively skewed meteoric calcite line, and late spars show depleted
δ 18 O values due to elevated temperatures during burial.
3. Aggrading neomorphism is common in original lime-mudstones and wackestones, producing pseudosparitic crystalline limestones.
4. Partial or complete dolomitization is widespread throughout the formation. The dolomitizing fluids were produced during gypsum precipitation and production of hypersaline brines in the basin. The seepage-reflux model best explains the dolomitization of the Minkinfjellet Formation.
5. Three mechanisms of silicification are common in the Minkinfjellet carbonates: (i) silica replacement of gypsum nodules, (ii) concretionary growth of silica nucleated by organic matter and (iii) replacement of calcite spar by megaquartz.
6. Quartz cementation by grain-contact dissolution and quartz overgrowth is the initial and dominant cement in Minkinfjellet Formation sandstones.
Pore-filling calcite spar post-dating quartz-cementation is abundant in many sandstones.
Acknowledgements .- This project was funded by Norsk Hydro ASA. We would specially like to thank Arve Lønøy at Norsk Hydro Research
Centre and Tommy Samuelsberg at Norsk Hydro Harstad for assistance and comments in the field and during analytical work. We also like to thank Mina Aase for all the help during the 2000 field season, and Eirik
Aaserød, Anne-Sofie van Cauwenberge and Mette Haugan Eliassen for useful field assistance during the 1999 and 2000 field seasons. Many thanks also go to Mateu Esteban for valuable comments during the
1998 field reconnaissance.
1. Three cement generations have been detected in the
Minkinfjellet Formation limestones: (i) Early marine fibrous cement, fringing individual ooid, peloid or skeletal grains; this cement is volumetrically speaking, of minor importance, (ii) drusy calcite spar precipitated during flushing by meteoric waters adjacent to former exposure surfaces, (iii) coarsely crystalline, blocky, pore-filling spar precipitated during deeper burial.
2. Oxygen and carbon stable isotope data from the three cement generations support the above inter-

330 NORWEGIAN JOURNAL OF GEOLOGY
Adams, J. E. & Rhodes, M. L. 1960: Dolomitization by seepage refluxion.
American Association of Petroleum Geologists Bulletin 44,
1912-1920.
Allan, J. R. & Matthews, R. K. 1982: Isotope signatures associated with early meteoric diagenesis.
Sedimentology 29, 797-817.
Baush, W. M. 1968: Clay content and calcite crystal size of limestones.
Sedimentology 10, 71-75.
Blount, D. N. & Moore, C. H. 1969: Depositional and non-depositional Carbonate Breccias, Chiantla Quadrangle, Guatemala.
Geological Society of America Bulletin 80, 429-442.
Bustillo, M. A., Delgado, A., Rey, J. & Ruiz-Ortiz, P. A. 1998: Meteoric water participation in the genesis of Jurassic cherts in the Subbetic of southern Spain - a significant indicator of penecontemporaneous emergence.
Sedimentary Geology 119, 85-102.
Chafetz, H. S. & Zhang, J. 1998: Authigenic euhedral megaquartz crystals in a Quaternary dolomite.
Journal of Sedimentary Research 68,
994-1000.
Choquette, P. W. & James, N. P. 1986: Diagenesis in Limestones - 3.
The deep burial Environment.
Geoscience Canada 14, 3-35.
Choquette, P. W. & Steinen, R. P. 1980: Missisippian non-supratidal dolomite, Ste. Genivieve Limestone, Illinois Basin: evidence for mixed-water dolomitization.
In Zenger D. H., Dunham J. B. &.
Ethington R. L. (eds.): Concepts and models of dolomitization, 163-
196. SEPM, Tulsa.
Cutbill, J. L. & Challinor, A. 1965: Revision of the stratigraphical scheme for the Carboniferous and Permian rocks of Spitsbergen and Bjørnøya.
Geological Magazine 102, 418-439.
Dallmann, W. K., Gjeldberg, J. G, Harland ,W. B, Johannesen, E. P.,
Keilen, H. B., Lønøy, A., Nilsson, I. & Worsley, D. 1999. Upper Paleozoic lithostratigraphy.
In Dallmann, W. K. (Ed.): Lithostratigraphic
Lexicon of Svalbard. Review and recommendations for nomenclature use. Upper Paleozoic to Quaternary Bedrock, 25-126. Norsk Polarinstitutt, Tromsø.
Eliassen, A. 2002: Sedimentological and diagenetic investigations of the Minkinfjellet Formation, central Spitsbergen, Svalbard; with particular emphasis on the origin and evolution of a large paleokarst system.
Unpublished dr.scient. thesis, University of Bergen,
127 pp.
Fallick, A. E. & Hamilton, P. J. 1989: The isotopic geochemistry of ocean waters through time. Transactions of the Royal Society of
Edinburgh: Earth Sciences 80, 177-181.
Folk, R. L. 1964: Some aspects of recrystallization of ancient limestones.
American Association of Petroleum Geologists Bulletin 48, 525.
Friedman, G. M. 1980: Dolomite is an evaporite mineral: evidence from the rock record and from sea-marginal ponds of the Red Sea.
In Zenger D. H., Dunham J. B. &. Ethington R. L (eds.): Concepts and models of dolomitization, 69-80. SEPM, Tulsa.
Gimenés-Montsant, J., Calvet, F. & Tucker, M. E. 1999: Silica diagenesis in Eocene shallow-water platform carbonates, southern Pyrenees.
Sedimentology 46, 969-984.
Given, R. K. & Lohmann, K. C. 1985: Derivation of the original isotopic composition of Permian marine cements.
Journal of Sedimentary Petrology 55, 430-439.
Gjelberg, J. G. & Steel, R. J. 1981: An outline of lower-middle Carboneiferous sedimentation on Svalbard: effects of tectonic, climatic and sea level changes in rift basin sequences.
In Kerr J. K., Fergusson A. J. & Machan L. C. (eds.): Geology of the North Atlantic Borderlands.
Canadian Society of Petroleum Geology Memoir 7, 543-
561.
Gregg, J. M. & Sibley, D. F. 1984: Epigenetic dolomitization and the origin of xenotopic dolomite texture.
Journal of Sedimentary Petrology 54, 908-931.
Johannessen, E. P. & Steel, R. J. 1992: Mid-Carboniferous extension and rift-infill sequences in the Billefjorden Trough, Svalbard.
Norsk
A. Eliassen & M. R. Talbot
Geologisk Tidsskrift 72, 35-48.
Karakitsios, V. & Pomoni-Papaioannou, F. 1998: Sedimentological study of the Triassic solution-collapse breccias of the Ionian zone
(NW Greece).
Carbonates & Evaporites 13, 207-218.
Kendall, A. C. & Harwood, G. M. 1996: Marine evaporites: arid shorelines and basins.
In Reading H. G. (ed.): Sedimentary Environments:
Processes, Facies and Stratigraphy, 281-324. Blackwell, Oxford.
Knauth, L. P. 1979: A model for the origin of chert in limestone.
Geology 7, 274-277.
Kupecz, J. A., Figuerosa, L., Aquino, R., Hernandez, E., Perez, M., Prieto, M., Salazar, R. & Hartmann, D. J. 1998: Characterization and reservoir quality assessment of a mixed carbonate siliciclastic reservoir; Cretaceous Escandalosa Formation, "O" member, Barinas-Apure Basin, Venezuela.
American Association of Petroleum
Geologists Bulletin 82, 1931.
Kwiatkowski, S. 1981: Formation of chert nodules controlled by the host deposit in the Carboniferous limestone near Debnik, Southern Poland.
Studia Geologica Polonica 68, 43-60.
Land, L. S. 1973: Holocene meteoric dolomitization of Pleistocene limestones, north Jamaica.
Sedimentology 20, 411-421.
Land, L. S. 1980: The isotopic and trace element geochemistry of dolomite:the state of the art.
In Zenger D. H., Dunham J. B. & Ethington R. L. (eds.): Concepts and models of dolomitization, 87-110.
SEPM, Tulsa.
Land, L. S., Salem, M. R. I. & Morrow, D. W. 1975: Paleohydrology of ancient dolomites: geochemical evidence.
American Association of
Petroleum Geologists Bulletin 59, 1602-1625.
Lohmann, K. C. 1988: Geochemical Patterns of meteoric diagenetic systems and their application to studies of Paleokarst.
In James N.
P. & Choquette P. W. (eds.): Paleokarst, 58-80. Springer Verlag,
Berlin.
Longman, M. W. 1977: Factors controlling the formation of microspar in the Bromide Formation.
Journal of Sedimentary Petrology 47,
347-350.
Lønøy, A. 1995: A mid-Carboniferous, carbonate-dominated platform, central Spitsbergen.
Norsk Geologisk Tidsskift 75, 48-63.
Makhnach, A. A. & Gulis, L. F. 1993: Chert nodules in Devonian and
Cretaceous carbonate rocks, Belarus. Article 3: origin.
Lithology and Mineral Resources 28, 504-513.
Maliva, R. G. 1987: Quartz Geodes: Early diagenetic silicified anhydrite nodules related to dolomitization.
Journal of Sedimentary
Petrology 57, 1054-1059.
Maliva, R. G. & Siever, R. 1989: Nodular chert formation in carbonate rocks.
Journal of Geology 97, 421-433.
Mattes, B. W. & Mountjoy, E. W. 1980: Burial dolomitization of the
Upper Devonian Miette Buildup, Jasper National Park, Alberta.
In
Zenger D. H., Dunham J. B. & Ethington R. L. (eds.): Concepts and models of dolomitization, 259-297. SEPM, Tulsa.
McCann, A. J. & Dallmann, W. K. 1996: Reactivation history of the long-lived Billefjorden Fault Zone in north-central Spitsbergen,
Svalbard.
Geological Magazine 133, 63-84.
McWhae, J. R. H. 1953: The Carboniferous breccias of Billefjorden,
Vestspitsbergen.
Geological Magazine 90, 287-298.
Milliken, K. L. 1979: The silicified evaporite syndrome - Two aspects of silicification history of former evaporite nodules from Southern
Kentucky and Northern Tennessee.
Journal of Sedimentary Petrology 49, 245-256.
Moore, C. H. J., Facundus, M. R. & Blount, D. N. 1969: Dedolomitization associated with evaporite solution.
Special Paper - Geological
Society of America, 205-206.
Morrow, D. W. 1982: Diagenesis 2. Dolomite - Part 2. Dolomitization models and ancient dolostones.
Geoscience Canada 9, 95-107.
Muchez, P. & Viaene, W. 1994: Dolomitization caused by water circulation near the mixing zone: an example from the Lower Viséan of the Campine Basin (northern Belgium).
In Purser B., Tucker M. &
Zenger D. (eds.): Dolomites: a volume in honour of Dolomieu, 155-

NORWEGIAN JOURNAL OF GEOLOGY
168. Blackwell, Oxford.
Pickard, N. A., Eilertsen, F., Hanken, N. M., Johansen, T. A., Lønøy, A.,
Nakrem, H. A., Nilsson, I., Samuelsberg, T. J. & Somerville, I. D.
1996: Stratigraphic framework of Upper Carboniferous (Moscovian-Kasimovian) strata in Bünsow Land, central Spitsbergen: palaeogeographic implications.
Norsk Geologisk Tidsskrift 76, 169-
185.
Pittman, J. S. & Folk, R. L. 1971: Length-slow chalcedony after sulphate evaporite minerals in sedimentary rocks.
Nature 230, 64-65.
Popp, B. N., Anderson, T. F. & Sandberg, P. A. 1986: Brachiopods as indicators of original isotopic compositions in some Paleozoic limestones.
Geological Society of America Bulletin 97, 1262- 1269.
Samuelsberg, T. & Pickard, N. A. 1999: Upper Carboniferous to Lower
Permian transgressive-regressive sequences of central Spitsbergen,
Arctic Norway.
Geological Journal 34, 393-411.
Steel, R. & Worsley, D. 1984: Svalbard´s post-Caledonian strata - an atlas of sedimentational patterns and palaeogeographic evolution.
In A. M. Spencer et al. (eds): Petroleum Geology of the North European Margin, 109-135. Graham & Trotman, London.
Tucker, M. E. 1976: Quartz replaced anhydrite nodules (Bristol Diamonds) from the Triassic of the Bristol District.
Geological Magazine 113, 569-574.
Tucker, M. E. 1991: Sedimentary Petrology.
Blackwell Scientific Publications, 260 p.
Tucker, M. E. & Wright, V. P. 1990: Carbonate Sedimentology.
Blackwell
Scientific Publications, 482 p.
Veizer, J., Fritz, P. & Jones, B. 1986: Geochemistry of brachiopods:
Oxygen and carbon isotopic records of Paleozoic oceans.
Geochimica et Cosmochimica Acta 50, 1679-1696.
Veizer, J., Bruckschen, P., Pawellek, F., Diener, A., Podlaha, O. G., Carden, G. A. F., Jasper, T., Korte, C., Strauss, H., Azmy, K. & Ala, D.
1997: Oxygen isotope evolution of Phanerozoic seawater.
Palaeogeography, Palaeoclimatology, Palaeoecology 132, 159-172.
Diagenesis of the Minkinfjellet Formation, Central Spitsbergen 331