The cementation zone forms an envelope of variable width along the Dombjerg Fault and, based on investigation of additional outcrops in the 2018 field season, we estimate its maximum width to be ∼ 1.5 km (Fig. 2), which is larger than previously estimated by Kristensen et al. (2016; i.e., ∼ 1 km). Within this zone, calcite cementation is prominently distributed and occurs in sedimentary layers of all grain sizes, while uncemented layers occur occasionally and are usually less than 50 cm thick. At the fault-distal limit of the cementation zone, cementation is mainly confined to conglomerate beds and outcrops show cemented layers with “holes” formerly filled with now-eroded uncemented material (Fig. 3a). It appears that the cemented layers enclosed pockets of calcite-cement-free sediments. Farther to the east, calcite cement is absent in the vast majority of outcropping sediments (Fig. 3c) and only occurs occasionally as single isolated lenses in predominantly fine to coarse sand matrix-supported conglomerates.

Calcite cement occurs as micrometer-sized spar, anhedral equant spar with crystal sizes up to 200 µm, and subhedral poikilotopic calcite with crystal sizes up to 800 µm (Fig. 4c–f). Micro-spar is present in 12 of 21 samples, all of which also comprise a varying amount of equant calcite spar. The latter is dominant in four samples. Poikilotopic texture occurs in only two samples from within the cementation zone (samples G-9, TKB2) and in three samples from cemented lenses outside the cementation zone (sample H-5). Notably, sample TBK2 hosting poikilotopic calcite derives from the same bed as sample TBK1 hosting dominantly micro-spar (samples taken ∼ 1 m apart from each other; Fig. 4c, d).

Other notable diagenetic features are the following:

Feldspar grain dissolution appears in varying degrees and is generally more dominant in the cemented samples, where it is partially replaced by calcite cement. Feldspar overgrowth predates calcite cementation and occurs in both cemented and uncemented samples (Fig. 4b, h).

In all outcrops within the cementation zone, calcite veins were found (Fig. 3d), which have an overall NNE to NE trend (67 veins measured in total; Fig. 2). Outcrop surfaces were exposed in 3-D, and we hence conclude that this dominant vein trend is true and not biased by outcrop orientation. Veins exclusively occur in cemented parts of the sandstone and are approximately equally distributed in outcrops with spacing varying in the decimeter to meter scale and do not form clusters. Single vein thicknesses span from <1 to 70 mm and veins appear as mode I fractures; i.e., possible shear displacement is only observed in three of 14 samples. In all cases, the veins have a sharp contact with the wall rock, i.e., wall rock grains are cut by the fracture (Fig. 7). All analyzed veins exhibit at least one generation of elongate to blocky syntaxial crystal growth. Two-thirds of the analyzed veins comprise two or more vein growth generations, indicated by either crack-seal texture (i.e., precipitation after reopening of fracture) or growth zonation, as visible in thin section view (Fig. 7a, b). Calcite twin density is generally low and slip zones within veins are found in three samples.

Near the fault core within brecciated basement rock, two vein generations were identified: a syntaxial dolomite vein network that is cross-cut by a syntaxial calcite vein network (Fig. 3e). The dolomite network composes of continuous, anastomosing veins of thicknesses up to 30 mm wide (Figs. 3e, 7e). The calcite vein network consists of planar veins with thicknesses of ∼ 1–5 mm (Fig. 3e) and also envelopes dolomite in the center of dolomite-filled fractures (Fig. 7f). The dolomite veins are not displaced/offset by the calcite veins and the latter are not displaced by any other fracture set. In the transfer zone, connecting the Dombjerg and Thomsen Land faults (Fig. 2), only one calcite vein was found in a shear fracture that reactivated an older epidote vein situated in basement rock.

We determined formation ages of five cement samples and 12 vein samples within the sediments of the Lindemans Bugt Formation (Table 1, Fig. 8). As some veins comprise more than one growth generation, as evident by zoning or fracture re-opening (see Sect. 4.2), we obtained a total of 25 growth ages of the 12 vein samples. Age determination of veins within the basement and fault core were not successful due to unsuitable U and Pb signals.

Three cement ages fall within the previously reported depositional age of the Lindemans Bugt Formation (mid-Volgian–late Ryazanian; Surlyk, 2003) or close to its depositional age boundary with the Palnatokes Bjerg Formation (late Ryazanian–Hauterivian) with 150.6 ± 9.3 Ma (G-36cem), 143.7 ± 6.5 Ma (TBK1cem), and 139.4 ± 4.9 Ma (G-38cem), while two other cement ages are significantly younger with 103.3 ± 2.6 Ma (TBK2cem) and 104.1 ± 1.7 Ma (G-9cem). Of the vein samples, only one falls within the inferred rift stage with 139.3 ± 3.4 Ma (G-10 v1), while the majority of vein ages (both initial formation ages and vein reopening ages) fall within a range of ∼ 125–90 Ma (Fig. 8). Two vein ages are significantly younger with 50.1 ± 2.1 Ma (G-7) and 49.4 ± 2.1 Ma (G-2), closely postdating early Eocene plateau basalt extrusion and the onset of continental breakup (Fig. 8).

Clumped isotope temperatures of five pore-filling calcite cement samples were obtained. Temperatures from TBK1cem and TBK2cem, which derive from the same outcrop, are 42.0 ± 10.2 and 59.0 ± 9.1 ∘C, respectively. G-25cem, sampled at a similar distance to the main fault core, has a clumped isotope temperature of 42.4 ± 8.8 ∘C, G-36cem at the distal margin of the cementation zone has 56.6 ± 10.9 ∘C, and H-5cem, located well into the basin, has 44.5 ± 9.0 ∘C (Table 2, Fig. 2).

Clumped isotope temperatures of hanging wall veins fall into the range of 36.3 ± 9.4 ∘C (G-36 v1) and 77.6 ± 10.1 ∘C (G-22). In the two samples of which two vein generations were analyzed, a temperature increase exists from old to young vein phase (i.e., G-36 from 36.3 ± 9.4 to 58.2 ± 14.1 ∘C and G-10 from 40.3 ± 7.1 to 61.5 ± 10.5 ∘C; Table 2, Fig. 7a). However, taking all samples, including cement samples, into consideration, no clear relationship between calcite generation and temperature is visible (Fig. 9a), which is also true for the relationship of temperature and distance to the fault core (Fig. 9b). The clumped isotope temperature of the basement vein sample G-34, located in the transfer zone, yields 128.7 ± 19.1 ∘C. Near the fault core in the brecciated basement, the clumped isotope temperature of the calcite vein (TBK9cal) that cross-cuts the dolomite vein network falls with 68.8 ± 10.9 ∘C into the range of hanging wall vein temperatures. The older dolomite vein sample TBK9dol diverges from this pattern with a higher temperature of 106.5 ± 11.9 ∘C (Fig. 7f; Table 2).

Carbonate δ18OVPDB values range from -11.3 ± 0.09 ‰ to -6.1 ± 0.05 ‰ for cements and from -12.8 ± 0.05 ‰ to -4.1 ± 0.09 ‰ for most hanging wall veins, with the exception of vein sample G-7 with -21.4 ± 0.12 ‰ (Table 2). The basement vein sample G-34 comprises -18.0 ± 0.04 ‰ and the two fault core vein samples have values of -15.3 ± 0.33 ‰ (TBK9dol) and -11.7 ± 0.05 ‰ (TBK9cal).

The calculated fluid δ18OVSMOW values range from -4.3 ± 1.9 ‰ to -0.2 ± 1.6 ‰ for cements and -3.0 ± 1.7 ‰ to +1.9 ± 2.3 ‰ for hanging wall veins, while sample G-7 deviates significantly from this suite with -13.7 ± 1.5 ‰ (Fig. 9c; Table 2). The basement sample G-34 as well as the vein calcite sample TBK9cal also fall into this range with -0.4 ± 2.1 ‰ and -1.8 ± 1.7 ‰, respectively. The older vein dolomite sample TBK9dol from the fault core appears to have precipitated from an enriched fluid of +16.2 ± 1.6 ‰.

Carbonate δ13CVPDB values range from -18.2 ± 0.3 ‰ to -9.7 ± 0.8 ‰ for cements and -23.5 ± 0.5 ‰ to -11.3 ± 0.1 ‰ for hanging wall veins (Fig. 9c; Table 2). The basement and fault core samples differ significantly from these values with -5.5 ± 0.1 ‰ (G-34), -4.2 ± 0.1 ‰ (TBK9cal), and -2.2 ± 0.2 ‰ (TBK9dol). In general, and bearing in mind the limited database, samples in the hanging wall might show a slight trend from low carbonate δ13CVPDB and high fluid δ18OVSMOW values to higher carbonate δ13CVPDB and lower fluid δ18OVSMOW values (Fig. 9c). Also, a slight increase in fluid δ18OVSMOW and decrease in carbonate δ13CVPDB with increasing distance to the fault might exist (Fig. 9d, e).

For the minor element concentration analysis, we studied 10 calcite cement samples, 12 hanging wall calcite vein samples, two samples of veins from the fault core (one dolomite, one calcite vein), and one basement calcite vein sample (Table 3, Fig. 10). As some hanging wall samples comprised multiple vein generations, we measured in total 22 growth generations in the 12 hanging wall samples.

Common to all samples is a Sr concentration below or close to the detection limit and will therefore be neglected in the following. Cements comprise element concentration averages of 2429–11 862 (Fe), 2866–5045 (Mn), and 1152–7557 ppm (Mg) (Table 3). Hanging wall veins yield similar averages with values of 1436–13 598 (Fe), 2515–6719 (Mn), and 401–5180 ppm (Mg). The two Eocene-aged hanging wall veins (G-2 and G-7) differ significantly, with Fe and Mg concentrations being below the detection limit, and Mn concentration below detection in sample G-2, and 2390 ppm in sample G-7. The same applies to the basement sample G-34 with Fe, Mn, and Mg concentrations being below the detection limit. The fault core vein TBK9cal yields concentrations of 8342 (Fe), 1435 (Mn), and 887 ppm (Mg). Naturally, the older fault core dolomite vein TBK9dol diverges from the other samples with a high Mg and Fe concentration of 112 599 and 24 556 ppm, respectively, and a low Mn concentration of 1219 ppm.

The comparison of veins and respective cements from the immediate wall rock within a sample shows similar Mg and Mn concentrations for a majority of samples (i.e., samples G-4, G-9, G-22, G-25, and G-38; Fig. 10), which also partly accounts for Fe (G-4, G-25, G-38; Fig. 10). Overall, the Fe/Mn/Mg ratio is similar from cement to vein within half of the samples, whereas concentration and ratios differ from sample to sample (e.g., compare G-25 and G-38; Fig. 10). Vein generations within a sample do not show significant variations in minor element concentrations (Fig. S2 in the Supplement). No trends in concentration or ratio versus time or spatial distribution are evident.

Based on the U–Pb calcite cement ages of 150.6 ± 9.3, 143.7 ± 6.5, and 139.4 ± 4.9 Ma (samples G-36, TBK1, G-38), the formation of the cementation zone in the hanging wall has likely occurred during or immediately after the deposition of the host sediments of the Lindemans Bugt Formation during the rift climax (middle Volgian–late Ryazanian; ; Fig. 8). The 150.6 Ma age of G-36 is unrealistic as it predates the Lindemans Bugt Formation; however, its error margin reaches well into the time interval over which this formation was deposited Fig. 8; e.g.,. The upper formation age of the cementation zone is bounded by the age of vein G-10 v1 with 139.3 ± 3.4 Ma (i.e., close to the upper age boundary of the Lindemans Bugt Formation), as the rock had to be cemented to provide the tensile strength for the discrete fracture to form.

The two younger cement ages of 104.1 ± 1.7 (G-9cem) and 103.3 ± 2.6 Ma (TBK2) do not reflect a second cement growth phase but are interpreted to be recrystallization ages for the following reasons: in sample G-9, one vein generation with 115.5 ± 3.1 Ma (G-9 v4) is significantly older than the cement age. In line with our argument based on sample G-10 v1 above, it is plausible that the wall rock was cemented before ∼ 115.5 Ma to allow for a discrete fracture to form. In sample TBK2, derived from the same outcrop as TBK1, the cement has a well-developed equant sparitic to poikilotopic texture (Fig. 4d) with crystal sizes that exceed any other analyzed cement sample. This type of texture has elsewhere been interpreted as evidence for recrystallization e.g.,, which, in conjunction with the similar age of the G-9 cement, gives rise to our similar interpretation for the TBK2 cement, although we cannot fully exclude the possibility of a second cement growth phase.

Having established that the cementation zone formed immediately after deposition of the Lindemans Bugt Formation, its maximum formation depth is constrained by the thickness of the remaining part of the unit above the sampled intervals. The lithostratigraphic top of the Lindemans Bugt Formation is exposed on a ridge ∼ 4500 m away from the Dombjerg Fault (Fig. 2). Using a maximum depositional dip angle of 15∘ for fault-proximal Lindemans Bugt Formation sedimentary strata allows interpolation of the top of this formation towards the fault (Fig. 11). With this approach, we estimate a maximum thickness of the Lindemans Bugt Formation above sample TBK1cem of ∼ 1050 m and above sample G-36cem of ∼ 750 m. These thicknesses represent maximum estimates, because the depositional slope angle of the sediments decreases away from the fault. Applying an angle of 10∘ would, for example, lower the estimated thickness to 740 and 530 m above samples TBK1cem and G-36cem, respectively (Fig. 11). Hence, we estimate that the cementation zone formed at a burial depth of ∼ 1000 m or less (compaction not accounted for).

Formation temperatures of the cements can be assessed from the clumped isotope temperatures, although their interpretation needs to be taken with care: calcite is subject to solid-state reordering of C–O bonds at ambient temperatures >∼ 100 ∘C, which affects the Δ47 composition and subsequently deviates the clumped isotope temperature from the initial formation temperature . For dolomite, solid-state reordering starts at ambient temperatures of ∼ 150 ∘C . It is documented that this resetting is a challenge especially for samples from sedimentary basins, where the rock may have experienced high burial temperatures for long time periods . For the sedimentary rock of the Lindemans Bugt Formation, the absence of quartz overgrowth in both calcite-cemented and uncemented sediment samples provides a control on the maximum burial temperature. Temperature is a critical factor for the formation of quartz cement, which is known to start at ∼ 70 ∘C, and its growth rate increases significantly with increasing temperature e.g.,, while, e.g., silica supply is not regarded as a limiting factor e.g.,. Therefore, we argue that the analyzed samples from the Lindemans Bugt Formation have not been subject to temperatures above 100 ∘C and that the clumped isotope temperatures reflect the formation temperatures. However, this excludes samples that were subject to recrystallization, i.e., dissolution and reprecipitation , as well as the basement and fault core samples, whose age/temperature history is not known.

Despite being similar in age, the cement samples show a variation in formation temperatures, e.g., 42.0 ± 10.1 ∘C for TBK1cem, located ∼ 300 m away from the fault, and 56.6 ± 10.9 ∘C for G-36cem, located ∼ 1500 m away from the fault (Fig. 9a, b; Tables 1, 2). This indicates that fluid circulation and heat flow was heterogeneously distributed in the hanging wall and that highest temperatures are not necessarily found closest to the fault. This may root in permeability variations within the hanging wall deposits, due to, e.g., grain size, sorting, or onset of cementation, which may cause local perturbation or channeling of fluid flow and advective heat transfer. Subsequently, upwelling hot fluids along the fault would have adapted variably to the ambient geothermal gradient, depending on their fluid pathway and flow rate.

The calculated fluid δ18OVSMOW of the cements allow us to assess the fluid source from which the cements precipitated from. In general, the mean marine δ18OVSMOW is regarded as -1 ‰ for the Cretaceous e.g.,. However, it is shown that the marine δ18O has likely been inhomogeneous in the Early Cretaceous with reported values ranging from -5.3 ‰ to +1.5 ‰ , which reflects a similar spread as shown by modern marine δ18O . The fluid δ18OVSMOW values of the cements, spanning from -4.3 ± 1.9 ‰ to -0.2 ± 1.6 ‰ fit well into this range and likely reflects a marine or potentially mixed meteoric/marine fluid system, which is unsurprising given the marine deposition environment in the hanging wall basin. argue for the presence of a delta located to the north of the study area, between the Dombjerg and Thomsen Land faults (Fig. 1b), supplied by a large hinterland catchment area. This river system might have provided sufficient inflow of freshwater into the marine hanging wall basin to shift the δ18O signature to the more negative values. An additional possibility may be the local influx of meteoric groundwater: the δ18OVSMOW values of the cements appear to increase slightly with distance to the fault (Fig. 9d) and with the caveat that this observation is based on a limited number of data points, this may be taken to suggest that meteoric groundwater affected the proximal (near-fault) parts of the hanging wall but did not access the more proximal hanging wall deposits.

The analyzed veins cover a U–Pb age range from 140 to 90 Ma. Only one of these veins (sample G-10 v1) formed during the rift-climax stage (Table 1, Fig. 8), whereas the majority formed broadly between 125 and 100 Ma in the post-rift stage, accompanied by renewed fracture opening and vein formation between 115 and 90 Ma (Fig. 8). In the Wollaston Forland Basin, there is, to our knowledge, no published evidence for tectonic activity in Aptian to Turonian times. However, on northern Hold with Hope, ∼ 80 km south of our study area (Fig. 1), a rift period occurred from the late Valanginian to middle Albian times with faulting along a series of N–NNE-trending normal faults . Both vein ages and orientations in our study area fit well with these structures and might therefore correspond to this tectonic phase. On the rift counterpart across the Norwegian and Greenland seas, the Lofoten–Vesterålen segment of the Norwegian continental shelf, a rifting phase occurred during the late Albian and may have a minor contribution to fracturing or fracture reopening. We note, however, that with regard to the limited sample number, we cannot fully exclude the possibility of continuous vein formation from the syn-rift into the post-rift.

It is also possible that the veins formed in two stages by (i) fracturing during the main rift stage and (ii) calcite precipitation, in the post-rift stage. However, the vein generation G-10 v1 formed at the end of the rift climax (Fig. 8; Table 1), showing that the formation conditions were favorable for calcite precipitation in fractures. Hence, it would be reasonable to expect that also other fractures forming during rift climax should be calcite filled, and we therefore consider such a two-stage interpretation as less likely, and our preferred interpretation is that the veins formed in their entirety in the post-rift stage. The U–Pb ages of vein samples G-7 (50.1 ± 2.1 Ma) and G-2 (49.4 ± 2.1 Ma) diverge significantly from the main sample suit. These veins closely postdate the extrusion of the Eocene plateau basalts covering the Lower Cretaceous late syn-rift Palnatokes Bjerg Formation, which overlies the Lindemans Bugt Formation. This event is dated to 56–53 Ma and is related to the onset of continental breakup, which started at the latitude of the Wollaston Forland around 55 Ma . Hence, the formation of the veins could be directly related to deformation and uplift following breakup. A second striking difference to the main sample suit is the low fluid δ18OVSMOW value of -13 ± 1.5 ‰ of sample G-7. It falls well into the global δ18O range of -8 ‰ to -20 ‰ for modern meteoric water at 60∘ N , i.e., the paleolatitude of the study area at 50 Ma , and thus most likely indicates a meteoric fluid. We therefore interpret that deformation and, potential, rift shoulder uplift in response to continental breakup e.g.,, was responsible for vein formation, which at this time occurred under meteoric conditions.

As suggested above, the clumped isotope temperatures of the hanging wall veins most likely reflect their formation temperatures. These generally lie close to the cement formation temperatures, which may be explained by (i) similar formation depth, (ii) deeper formation depth but lower geothermal gradient, or (iii) perturbation of fluid flow and advective heat transfer. The latter argument should be valid as the cementation zone has likely created a setting for complex fluid pathways, whereas the evolution of the geothermal gradient is difficult to assess. The burial depth of the analyzed sedimentary rock has increased after the formation of the cementation zone due to deposition of the late syn-rift Palnatokes Bjerg Formation and Barremian to Albian post-rift sediments e.g.,. Two formation temperatures and ages are obtained of successive precipitation generations from two vein samples, respectively (samples G-10 and G-36; Table 2, Fig. 7a). Both samples yield an increase of formation temperature from old to young vein generation (i.e., from 40.3 ± 7.1 ∘C at 139.3 ± 3.4 Ma to 61.5 ± 10.5 ∘C at 90.4 ± 1.5 Ma in sample G-10 and from 36.3 ± 9.4 ∘C at 122.4 ± 1.3 Ma to 58.2 ± 14.1 ∘C at 112.0 ± 1.4 Ma in sample G-36), which might indeed reflect an increasing formation depth.

For the calcite and dolomite samples from the basement and fault core, it is unclear if the clumped isotope temperatures reflect the original formation temperatures, since an age control is missing. Therefore, these samples may have been partially reset due to ambient temperatures above the threshold of C–O bond reordering of ∼ 100 ∘C for calcite and ∼ 150 ∘C for dolomite in the course of time. The clumped isotope temperature of 128.7 ± 19.1 ∘C of the basement sample G-34 could therefore reflect the original formation temperature, a maximum ambient temperature, a cooling temperature if the ambient temperature was >100 ∘C but lower than the calcite formation temperature, or a mixture of these. For the fault core vein samples, only a partial heating signal may be present as the measured clumped isotope temperatures of both veins are still below the temperature threshold (i.e., a fully reset signal would give a clumped isotope temperature above 100 ∘C for calcite and 150 ∘C for dolomite). Therefore, we are confident that the original formation temperatures of the dolomite and calcite vein samples from the fault core are ≤106.5 ± 11.9 and ≤68.8 ± 10.9 ∘C, respectively. Since both vein networks have a continuous, non-faulted appearance in the outcrop (Fig. 3e), we argue that these veins formed in the late stage of fault activity.

One of the main questions around the formation of the cementation zone is the origin of calcium necessary for the calcite cementation. A first likely candidate for a source is calcium from seawater that circulated through the succession . Such a circulation may be expected to follow general subsurface flow patterns in coastal areas. In such areas, groundwater systems are commonly characterized by marine (i.e., seawater) groundwater circulation that is tidal, density, and thermal driven, and meteoric groundwater flow towards the ocean, driven by the hydraulic head e.g.,. In the presence of a continuous and undisturbed permeable rock, marine circulation is not restricted to the aquifer underneath the ocean but extends landward beyond the coastline underneath the body of meteoric groundwater (Fig. 12a). The landward extension of marine groundwater circulation partly depends on the permeability of the aquifer e.g.,. Along the Dombjerg Fault, which defined and delineated the coastline during the rift climax , the footwall is composed of low-permeable crystalline and metamorphic rock with a low fracture and vein density four joints plus one vein per meter;. On the other hand, the hanging wall succession has distinctly different flow properties than that of low-permeable crystalline basement. Where uncemented, these deposits are highly porous (Fig. 4g) and therefore presumably permeable, allowing for fluid flux at a much larger degree compared to low-fractured crystalline and metamorphic basement rock e.g.,. Hence, along the Dombjerg Fault, marine fluid circulation should be predominantly restricted to the permeable fault zone and hanging wall clastic succession (Fig. 12b). Upwelling of warm marine fluids along the fault may have been in a favorable condition for the precipitation of calcite from seawater within the vicinity of the fault zone.

A second candidate for a calcium source is Permian carbonate and evaporite deposits underlying the clastic hanging wall rift sediments. While a direct exposure of such rock along the Dombjerg Fault is missing, a Permian carbonate unit is located along the northern section of the Clavering Fault Fig. 1b;, which forms the southern extension of the Dombjerg Fault. Permian carbonate rock is also exposed at the southern crest of the hanging wall block of the Dombjerg Fault Fig. 1b;. It may therefore be a reasonable assumption that the Permian succession is also underlying the Lindemans Bugt Formation along the Dombjerg Fault, especially considering that the fault had been active since the Carboniferous . It has been proposed and modeled elsewhere that surface waters may convect thermally driven within faults e.g.,. Such convecting fluids within the Dombjerg Fault may have dissolved the carbonate and transported the solutes into the Lindemans Bugt Formation for the precipitation of calcite.

Other potential sources for calcium can be excluded: no carbonate rock is hosted by the basement and also carbonate deposits in the Palnatokes Bjerg Formation overlying the cementation zone-hosting Lindemans Bugt Formation i.e., Albrecht Bugt Member; can be ruled out as a source, since the cementation predates this deposition. Further, dissolution of biogenic carbonate clasts and feldspar alteration appear unlikely as a potential internal source. Although biogenic carbonate clasts occur sporadically in some outcrop sections and are visible in a number of thin sections, they do not show signs of dissolution or recrystallization. The degree of feldspar alteration in the sandstone is low and hence will not have provided a significant amount of calcium. Therefore, whether calcium from seawater or dissolved underlying carbonates, it is clear that the calcium must have been introduced into the sediments of the Lindemans Bugt Formation by advective transfer.

CO2 that is acquired for the formation of calcite may have been generated and transported along in the same process. Yet, the negative δ13CVPDB values of the cements hint towards an (additional) internal source: δ13C values of calcite cements range from -18.2 ‰ to -9.7 ‰, which are typical values for CO2 deriving from the degradation of organic matter e.g.,. The high uranium content of the cements, which allowed the U–Pb calcite dating, also indicates degradation of organic matter, which is commonly bonding uranium e.g.,. Organic matter is common in the Lindemans Bugt Formation, composing mostly of ammonites, bivalves, belemnites, and transported plant and wood fragments , and the presence of pyrite in the sediment points towards an organic matter degradation in the sulfate reduction zone e.g.,.

Fe and Mn incorporated in the calcite cements and veins are most likely sourced from within the sediments, as their concentration in marine water is generally very low e.g.,. Notable is the overall high Fe concentration in the samples, while at the same time being highly variable across the sample suit (Fig. 10, Table 3). A major Fe source has likely been biotite, which is frequent in the sandstone, and the formation of pyrite along these grains demonstrates that Fe was released from biotite (Fig. 5). An inhomogeneous distribution of biotite in the sandstone might subsequently be responsible for the spatial variability of Fe concentration in the calcite. The low Sr concentration is more difficult to explain: Sr uptake into calcite is strongly temperature and precipitation rate dependent, i.e., the Sr concentration in calcite decreases with increasing temperature and increases with precipitation rate e.g.,. The low Sr concentration may therefore be a result of the elevated precipitation temperature in reference to calcite precipitating at or near the seafloor.

Throughout its seismic activity, the Dombjerg Fault was permeable due to repeated fracturing, as commonly assigned to fault zones e.g.,. The anastomosing dolomite vein network (represented by sample TBK9dol) close to the fault core suggests the fault represented a conduit for vertical (up-fault) fluid migration. It yields a fluid δ18OVSMOW value of +16.1 ± 1.6 (Table 2), which is by far the highest value determined in our sample suit and is indicative for a metamorphic fluid that was likely sourced from greater depths. On the contrary, its formation temperature of ≤106.5 ± 11.9 ∘C is not particularly high, which might reflect a moderate flow rate allowing the fluid to adapt to the ambient wall rock temperature.

The younger calcite vein network cutting through the dolomite veins is less pronounced and consists of thinner veins. Its fluid δ18OVSMOW value of -1.8 ± 1.7 ‰ (sample TBK9cal; Table 2) presumably indicates a marine fluid, which may be taken to suggest that the marine groundwater circulation was responsible for its formation. The δ13C value of -4.2 ± 0.1 ‰ of this vein differs significantly from hanging wall cements and veins (-23.5 ‰ to -9.7 ‰). Also, the minor element concentration ratio diverges from hanging wall samples with very low Mn and Mg concentrations (Table 3, Fig. 10). This might indicate that the marine groundwater circulation, from which the calcite in the fault core fracture network precipitated, was decoupled from groundwater circulation in the hanging wall sediments.

In the hanging wall, the cementation zone will have, once formed, significantly lowered the permeability of the fault-proximal hanging wall sediments, as the carbonate cement reduces the pore space towards 0 %. During progressive formation of the cementation zone and accompanying matrix porosity and permeability loss, fluid circulation becomes increasingly dependent on flowing through fractures developed within this zone. Here, the magnitude of circulation and flow direction is then guided by the connectivity and orientation of the fracture network.

The minor element concentration of cements and veins (Fig. 10) may shed light on the fluid flow rate and degree of fracture connectivity in the cementation zone. There is a striking similarity of minor element concentrations between veins and the respective wall rock cement in around half of the analyzed samples, while being variable from sample to sample (Fig. 10, Fig. S2 in the Supplement). Noteworthy is the compositional difference between samples G-36 and G-38: these samples are located ∼ 2 m apart from each other in the same outcrop, yet especially the Fe concentration of cements and veins varies between the samples, while being similar from respective wall rock cement to vein calcite. While sample G-36 yields an Fe concentration below average, sample G-38 yields the highest Fe concentration of all samples (Fig. 5, Table 3). The analyzed sample (G-38) hosts a very large quantity of biotite (Fig. 5), which serves as a Fe source, and could explain this high concentration. If an advective fluid was responsible for both the precipitation of cement and vein calcite, it must have kept a stable minor element composition over ∼ 20 Myr (i.e., the age difference between wall rock cement and vein calcite), which at the same time had to remain highly variable on a local spatial scale. We regard this as unlikely and therefore argue that the solute which formed the vein calcite likely was derived from diffusion from the calcite cement into the fracture. Diffusion of mass from local wall rock into a fracture is a common source of vein material and is promoted by, e.g., chemical and pressure gradients between wall rock and fluid-filled fracture (e.g., , and references therein; , and references therein). In addition, calcite is highly susceptible to pressure solution e.g.,, which may further stimulate the availability of diffusive material. It should be noted that the possibility of a later equilibrium between cements and vein through solid state diffusion cannot be fully neglected, as this is a process that is poorly understood at low temperatures, though experiments point towards an unlikely mechanism .

If true, i.e., diffusion being responsible for the vein calcite formation, this indicates very slow advective fluid circulation. Hence, despite being fractured, the effective permeability of the cementation zone remained low, which may root in poor fracture connectivity. The latter may be a reflection of the predominant parallel-striking veins and their near-vertical dips, causing limited fracture intersections (Fig. 2).