Author:

Keywords:

Travertine, Reservoir, Analogue, Facies, Geology, Sedimentology

Abstract:

Travertines are continental carbonates that precipitate due to CO2 degassing of supersaturated spring waters and are typically characterised by their eye-catching porosity. In addition to the more conventional brackish stromatolites, they have been envisaged as potential Pre-Salt reservoir rock in the South Atlantic. Studying outcrop analogues helps to better understand heterogeneities, lateral variability and properties of reservoirs. The Denizli Basin in the West Anatolian Extensional Province in western Turkey is well-known for its numerous travertine occurrences. An integrated sedimentological, diagenetical, geochemical and petrophysical analogue investigation was executed on the Kıllık dome travertines of the Ballık area, the largest travertine site in the Denizli Basin. Industry-formulated challenges were studied with an independent academic approach. Research questions, like the distribution and extension of the different geobodies and their petrophysical properties, are formulated in a reservoir analogue context to better understand travertine development. Depending on the scale of the observations, results and heterogeneities, they can be translated in terms of application for exploration or production, or both. The main research aims are defined as: Reconstruction of the 3D geo‑architecture of the deposits Characterization of the different lithofacies - lithotypes – lithofabrics Reconstruction of the depositional history, CO2-origin and deducing the source rocks of the fluids Delineation of diagenetic features and their influence Reconstruction of the palaeo-fluid characteristics, in particular the travertine precipitation temperature Although these aims are formulated as individual points, they are strongly intertwined. The reconstruction of a 3D geo‑model is combined with a detailed sedimentological description from fabric, to lithotype, lithofacies and geobody scale. The observations, descriptions and reconstructions of the sedimentological structures on the different scales form the starting points for the depositional reconstruction. Sedimentological interpretations can only be executed when also the diagenetic overprint is understood. And ultimately, it is the fluids and their properties that decide when, where and what is formed both during deposition and diagenesis. Such an integrated study can only be successful when a multi-method framework is applied within an upscaling philosophy, from the field to the lab. The first and second aims (Aims 1 & 2) of the study included the reconstruction of a three-dimensional geo-model and a detailed sedimentological description from fabric to lithotype, lithofacies and geobody scale, with a focus on integrating pore-types. 3D modelling was combined with field and microscopy observations. The quarry outcrops were reconstructed in 3D based on three different methods, namely by LiDaR*, photogrammetry and with SketchUp. Point clouds are more complex for facies delineation and visualisation. Consequently, in this study the simplified quarry model in SketchUp was used to easily and quickly draw complex geobody edges based on detailed line drawings. Interpolation between the Ece and Faber quarries was executed in SketchUp based on occurrence, distribution and slope, as well as nature of the facies based on observations in active settings and literature. The large-scale depositional Kıllık dome travertine system can best be compared to the mound/fissure ridge complex development as described by Pentecost (2005), with the spring locations controlled by the tectonic activity within the Denizli Basin (Van Noten et al., 2013). However, the exact position of the main fluid vent was not found inside or near the quarries. Based on the northwest-dipping slopes of the deposits, the main source should be located southeast of the studied quarries, near the centre of the domal area. The lateral extension of the sub-horizontal travertine facies suggests at least one additional major source at the time of travertine precipitation. Peloidal, phyto and dendritic lithotypes dominate the studied travertines. The presence of honeycomb structures, bacteriform shapes and encrusted bacterial or fungal filaments related to their fabrics suggest a microbial influence. The travertine build-up can be split up into four main systems (Chapter 2). The first system, represented by the sub-horizontal and biostromal reed travertines, formed in a shallow dominantly sub-aquatic environment. The second system consists of the non-travertine intercalation while the third system consists again of travertine, mainly represented by the sloping facies, formed in a thin water film in a dominantly sub-aerial setting. Finally the fourth system reflects the levelling up of the travertine system with a gradual transition to dominantly marly lacustrine strata. The sub-horizontal facies show an aggradational stacking pattern. A general progradation of travertine development is apparent based on the occurrence of stacked waterfall travertines. The progradation results in sigmoidal clinoforms-like deposits with downlap terminations, mainly against the laterally occurring marl-conglomerate deposits. Next to the sedimentological SketchUp model, the geo-architecture was also reconstructed based on variogram analyses of the facies followed by kriging in SGeMS. The geostatistical analysis started from pseudo-logs* (Chapter 7.1), an approach that can be applied to reservoirs. Transition probability calculations and indicator kriging have shown that facies type occurrences are non-random and depend on the preceding facies. The cascade facies dominates and occurs together with the waterfall facies in the middle part of the quarries. The sub-horizontal facies is found in the bottom part of the quarries and the reed facies is mostly found on top of the cascade and waterfall facies. Analogue studies have the advantage of being able to gather a much larger, more detailed, dataset, enabling a more accurate simulation of the facies probabilities (Chapter 7.2). In the direct proximity of the quarry walls, the kriged results are well in accordance with the expectations and the sedimentological model. However, more distal from the sampling grid, there are many inconsistencies. This is mainly because kriging can be considered an inter- rather than extrapolation method, leading to the overestimation of the dominant facies, or a high uncertainty of the facies. As an alternative method, multiple point geostatistics should be considered, where via a training image the sedimentological relations can be incorporated. Characterisation of the travertines and the different facies (Aim 2) was also based on geochemistry and petrophysical properties. Elemental geochemistry analyses were executed by Inductively Coupled plasma atomic emission spectroscopy (ICP-OES) and were statistically processed in R. The petrophysical analyses include standard helium porosimetry* and specific gas permeability* measurements. The travertine lithofacies are associated with different pore types. The travertine samples are heterogeneous and exhibit large-scale ranges for both porosity (3.3 – 46 %) and permeability (0 – 14600 mD). Travertine lithofacies exert significant control on porosity and permeability. The highest porosities and permeabilities are found in the waterfall facies, with framework and mouldic porosity. The lowest porosities and permeabilities are found in the sub-horizontal facies with layer parallel pore types. The results of the geochemical analyses proved to depend strongly on the applied methodology. Clays, oxides and organic matter partly dissolve and exchangeable ions will contribute to the elemental concentrations. The elemental signature of the samples will thus largely depend on the non-carbonate input into the system and thus mainly their lithology. Distinction for most travertine facies is not statistically significant. The use of diluted weak acids minimizes but does not exclude this problem. Even for sequential extraction procedures digestion of non-carbonate phases is inevitable. Applying different digestion methods allows linking elements to their chemical phase. Statistical processing of the data enables phase separation for each digestion method individually. Elements of continental carbonates can thus be linked to their incorporation site and their origin. Carbonate phase related elements (e.g. Ca, Mg, Sr and S) provide valuable information in deducing the fluid source rock. Non-carbonate fraction related element geochemistry (e.g. Al, Fe, Ti and Si) can be used as an important additional tool to evaluate detrital contamination and sample suitability for U/Th-dating. For the fourth aim (Aim 4), i.e. the distinction between depositional and diagenetic fabrics, field observations were mainly combined with those from optical, fluorescence, cathodoluminescence and scanning electron microscopic petrography, and results from elemental and stable isotope geochemistry*. Separating secondary from primary fabrics has always been a challenge in continental carbonate research. Macro-scale diagenetic alteration of the travertines was limited, as inferred from the lack of extensive cementation or dissolution. On micro-scale, the Ballık travertine fabrics usually show no cathodoluminescence. Only late diagenetic cements, for example along fractures and faults seem to show a weak and sometimes somewhat brighter cathodoluminescence (CL). Remarkably, also cloudy micrite and certain sparite zones in stalactite-like features show CL. The cloudy micrite is found along non-luminescent opaque micrite. Closing of the internal part of the stalactite-like system provided a reference point in the paragenesis of the travertine in the core of the structures, that likely can be extrapolated based on fabric similarities to the whole travertine body. In this way, cloudy fabrics associated to primary organic rich micrites in the whole travertine body can be interpreted as secondary. The studied stalactite-like features thus possibly act as a key in unlocking the early diagenesis of the travertines. The open-system behaviour of the travertines has implications for diagenetic and other geochemical studies, especially Uranium-Thorium analyses. The third and fifth main aims (Aims 3 & 5) include the reconstruction of the depositional environment, delineation of the CO2-origin, characteristics and source rocks of the ancient travertine precipitating waters. For these aims, in addition to the previous mentioned methods, also results from radiogenic strontium isotopes*, microthermometry*, crush-leach* and clumped isotopes* analyses were included. Geochemistry of the travertines and fluid inclusions reflects the precipitation fluid composition, revealing important information on the large scale hydrology. The main travertine precipitating and cementing fluids are related to a low-salinity meteoric fluid with variable temperatures dominantly below 50°C. For the most extreme measured, calculated and reported values of δ18Owater, the calculated temperatures based on the main δ18Otravertine cluster range reach up to 65°C, with the most likely precipitation temperatures between 22 and 46°C. In contrast, a later phase of travertine cementation is related to a fluid with a higher salinity. Strontium and stable oxygen-carbon isotope signatures of the main travertines point to a mixing mechanism of palaeofluids with deep originated heavy carbon CO2 and lighter carbon CO2 of a shallow origin. Group I elements (Ba, Ca, Mg, Na, S and Sr) in the travertines mainly depend on the concentrations in, and the interaction intensity and duration with, the element source rocks and the distance to the vent. Based on the strontium and stable isotope signature, the elemental characterization of the travertines and fluid inclusions, the salinity of the fluids and the geological proximity, the Triassic Lycian Nappes acted as the main parent carbonate rock. The signature of the Kaleboğazı Formation was reflected in the geochemistry. The Ballık-Belevi travertines can thus be considered as endogenic travertines, in which regardless of the water temperature, a deeply-derived significant water component is involved in precipitation. The water compositions are most similar to those reported at the (nearby, along the same fault system) Karahayit and Pamukkale hot springs. In order to determine the evolution of the fluid through time (Aim 5), dating is necessary. From a hydrocarbon reservoir analogue point of view, the timing of active precipitation and associated environmental conditions (Aim 3) are also key and correlation horizons are of interest (Aim 1). The precipitation rate of continental carbonates, including travertines, is also of high importance (Aim 3), making dating critical. A uranium-thorium analysis*, the most applied method for fossil travertine dating, was attempted on the Ballık travertines. The resulting ages are close to the limit of the dating method. Even with corrections for detrital contamination, ages are out of stratigraphic order. The diagenetic susceptibility of the deposits poses problems in terms of accurate and precise dating. The results are interpreted as indicating open-system behaviour on primary deposits that precipitated >550 ka ago. Effort should be put in alternative dating methods. This study characterised travertines as a potential reservoir analogue for the Pre-Salt reservoir carbonates. Within the scope of the postulated research aims, the Ballık travertines are results are considered rather independent from the Pre-Salt observations. The question raised, however, how the results can be coupled back to the Pre-Salt setting. Information from the Pre-Salt reservoir carbonates is based on material from cores and seismics. On core-scale, the interest from industry with regard to shrub structures became clear. Consequently, shrub structures from the travertines were compared with publicly available petrographical and petrophysical data of those from the Pre-Salt. New classifications for shrub-related lithotypes and associated pore types are presented. Reported Pre-Salt fabric and morphology types, were recognized in the heterogeneous Turkish travertine shrubs. For both the Ballık-Belevi and Pre-Salt fabrics a microbial influence is suggested and a fibrous nature of the fabrics is observed. The Ballık-Belevi travertine shrubs in general show a higher presence of micrite fabrics and their porosity (3-15 %) and permeability (0-505 mD) are usually lower than for reported Pre-Salt counterparts. The highest porosities and permeabilities are measured for horizontal samples dominated by growth framework porosity. The results of this analogue study are mainly on a sub-seismic scale. However, the presented lithofacies geobody model, in combination with petrophysical and acoustic data, enable a seismic-scale reservoir analogue modelling. From the synthetic seismic* results, it is clear that the external boundaries of a travertine dome in the Pre-Salt setting will result in strong reflectors. For the depth of the Pre-Salt reservoirs, a resolution in the order of 30 m was obtained. This implies that the internal dome morphologies cannot be deduced. For prospection based on seismic data, an external mound or domal morphology, associated to normal faults in an extensional setting, possibly associated with a positive antecedent morphology and onlapping lithologies on the side of the dome forms a potential travertine occurrence. Based on evaluation of seismic lines, the size, morphology and extensional setting of the Kıllık dome are in agreement with the aggradational-progradational carbonate build-ups reported for the Pre-Salt play of the South Atlantic. The combination of seismic data with core data, however, remains of vital importance. This study provides a unique three-dimensional insight into the Ballık travertine architecture that potentially can be used as an analogue for subsurface travertine reservoirs worldwide. Facies occurrence, thickness and lateral continuity will have a major impact on porosity and permeability distributions. The geostatistical modelling is an effective way for facies, porosity, and permeability prediction at unsampled locations. With the incorporation of statistics, an upscaling from log section measurements to a 3D porosity, permeability and facies distribution model can be achieved on reservoir scale (km²). This study also illustrates the importance of the combined use of δ13C, 87Sr/86Sr signatures and elemental analyses in the delineation of the CO2-origin and fluid source rocks of travertine precipitating waters. The study highlights the novel application of fluid inclusion research in combination with clumped isotopes and crush-leach analyses in unravelling the precipitation temperature and precipitation fluid characteristics of continental carbonates. The Ballık-Belevi travertines show highly complex pore networks, as observed from Computed Tomography images, making them of interest for the development and optimization of pore characterization methodologies and techniques, following an upscaling approach. The presented shrub and pore-type classification can be applied to shrubs worldwide. *the measurements for these methods were externally delegated (specified in the methodology section of the related chapters).