Depositional control on diagenetically-induced heterogeneity in triassic red beds sandstones

Resumen

Analysis of the spatial and temporal evolution of diagenetic features is of paramount importance for the evaluation and modelling of petrophysical parameter distribution in siliciclastic reservoir rocks. The recent proliferation of software for reservoir quality prediction and modelling has changed the approach of the industry to the appraisal and recovery phases of hydrocarbons exploration and production. Most of these software focus on specific aspects as for instance quartz cementation modelling which is much more easily predictable due to the chemical conditions of the diagenetic environment in which it develops. A completely different task could be to forecast the occurrence and amount of clay minerals and their impact on reservoir quality by understanding their distribution within a depositional system and their subsequent post-depositional modifications. Diagenetic patterns are linked to an array of depositional-dependant (depositional fabric, pore-water chemistry, intrabasinal components and bioturbation) and independent (burial-thermal evolution of the basin, residence time in specific P-T conditions) factors which control the further sediment modification, from shallow (early) to deeper burial conditions. These early mechanical and chemical processes proceed along different pathways directly related to depositionally-governed differences in texture and composition. Mutual interaction of depositional features and sandstone framework composition with diagenetic processes and fluids determine the compaction intensity, the amount and types of cements, the extent of dissolution and, therefore, the amount of residual primary porosity or generation of secondary porosity, so eventually permeability and reservoir quality. All these factors (e.g. depositional environment, provenance,climate, tectonic) are intimately connected and represent – among others – the necessary input parameters to predict and model reservoir quality. However, in many cases information about depositional facies or environment is not fully captured from subsurface core samples. It is therefore of key importance the use of outcrop reservoir analogue studies which provide quantitative, geo-referenced datasets that serve as input in state-of-the-art forward modelling software of sandstone diagenesis and petrophysical properties. The aforementioned facts constitute the starting point of this PhD project and the rationale behind the selection of both the study areas and the analytical approach. The aim of this Thesis is to reduce the uncertainty on RQ modelling through a multidisciplinary approach integrating the study of depositional facies and the analysis of early diagenetic process in reservoir-analogue outcrops and behind-outcrop core data analysis. The latter is of particular interest for the hydrocarbon industry as it remarkably enhances the relatively limited interpretations obtained from traditional core description by direct outcrop-faced contrast. The study areas have been selected according to the relevance of both (i) the represented sedimentary environments of interest for the hydrocarbon exploration and (ii) features of the outcrops in terms of reservoir analogues. The selected fluvial depositional environments are exposed in the Triassic succession of Central SE Spain (herein referred as TIBEM) and in the Argana Basin (Western High Atlas, SW Morocco) and correspond to: braidplain and floodplain (including overbank and channelized deposits) environments in Spanish area; and braided and straight fluvial systems in Moroccan area. Due to formation geological conditions, such areas are ideal outcrop analogues of a number of well-known Triassic reservoirs including the TAGI (Trias Argilo-Gréseux Inférieur) reservoirs of North Africa and the Bay of Fundy (Nova Scotia, Canada). The methodology adopted for the development of this research encompasses: a thorough selection of samples according to pre-existing and newly acquired sedimentological criteria from outcrop and behind-outcrop core analysis; high-resolution petrological analysis following the up-to-date techniques (e.g. optical and electron microscopes, XRD, cathodoluminescence, microprobe) and developing done-on-purpose characterization methodologies; determination of the main petrophysical parameters (porosity, permeability and pore size distribution) by mercury injection-capillary pressure. The first part of this research has been focused on the Triassic succession from Central SE Spain aiming at (i) recognising vertical compositional and textural changes in sandstone samples from the fluvial deposits of floodplain environment of Sequence II to the braided environment of Sequence IV and (ii) evaluating the impact of such compositional and textural features on diagenetic processes and pore network evolution to (iii) eventually classify the studied sedimentary environments according to their potential as reservoir rocks. Obtained results document a significant maturity increase up-section from floodplain to braidplain environment marked by changes in depositional texture and detrital composition. Petrographic evidences suggest that porosity loss happened during the very early diagenetic stages through different diagenetic processes according to facies. In floodplain environment, overbank sandstone deposits have been mainly affected by mechanical compaction and gypsum cementation. The channelized deposits from floodplain include meandering and straight fluvial systems and display intermediate mechanical compaction degree and variable amount and type of dominant authigenic phases depending on targeted depositional facies. In the braidplain environment, primary porosity is still well-preserved with K-feldspar overgrowths as the prevailing cementing phase. Occurrence of gypsum-cemented slough channel facies and carbonate-cemented post-sedimentary fractures, where porosity drastically drops, in the braidplain, deteriorates the potential of this depositional environment as reservoir rock. The link between depositionally-controlled parameters (i.e. textural features and abundance of intrabasinal components) and spatial and temporal distribution of diagenetic processes has been further investigated in meandering channel deposits from Sequence II (Central SE Spain) and braided and straight fluvial deposits from the Argana Basin (Western High Atlas, SE Morocco). In the meandering example, the study has been performed on core samples from 4 behind-outcrop wells targeting the main depositional facies: channel, point bar, scroll bar and chute channel. Besides to traditional petrographic point count, a done-on-purpose petrographic semi-quantitative characterization of pore spaces has been developed. The statistical integration of petrographic and petrophysical data revealed a major role played by hydraulic sorting processes and its effects on the diagenetic evolution and reservoir properties. Obtained results show the direct relationship between depositional facies and the abundance and distribution of detrital matrix, which favours mechanical compaction but inhibits early cement precipitation. Multivariate statistical processing of the integrated data demonstrates that large (>1 µm) and well-connected primary intergranular pores are the main contributors to permeability in the more heterogeneous samples. Secondary porosity seems to be mostly related to smaller (<1 µm) and randomly distributed dissolution pores. Sedimentological, petrographic and petrophysical data have been integrated to calculate the Reservoir Quality Index which was then used as input for a deterministic facies 3D PETREL® model. Remarkably, the RQI model works perfectly, enhancing the relationships between reservoir properties and facies distribution and identifying both the channel margin and the scroll bar as the depositional facies having the best reservoir quality. The obtained model can be therefore used to identify potential preferential-flow pathways and be successfully applied to Enhanced Oil Recovery (EOR) strategies (water or steam injection). In braided and straight fluvial deposits from the Argana Basin, occurrence and amount of different types of rip-clasts have a direct and substantial control on early diagenetic evolution and subsequent reservoir deterioration. A done-on-purpose classification of these grains according to their composition and mechanical behaviour leads to three main types, namely: muddy rip-up clasts; dolomitic muddy rip-up clasts and dolomite crystalline rip-up clasts. Muddy rip-up clasts behave as ductile grains and their presence in sandstone framework results in a significant early reduction of primary porosity and irreversible loss of IGV. Conversely, dolomite crystalline rip-up clasts act as rigid grains against mechanical compaction but favour precipitation of early pervasive dolomite cement that occlude primary porosity and preserves IGV. At pore- and depositional-scales, spatial distribution of rip-up clasts in continuous layers (e.g. muddy rip-up clasts) or in specific depositional facies (e.g. dolomite crystalline rip-up clasts) and their associated diagenetic processes may impact reservoir quality by generation of vertical and 3D fluid flow barriers and baffles that compartmentalize the reservoir. Obtained quantitative estimation of early reduction of primary porosity by compaction and cementation on the basis of the abundance of these type of grains can provide useful data for the better understanding of reservoir quality in analogous basins. An additional case study is presented as part of this Thesis, following the same approach about linking depositional features to early diagenetic processes in order to predict their impact on reservoir quality. This study was encouraged by the need of understanding the low-than-expected flow rates occurring at a geothermal well from the aeolian facies of the Rotliegend geothermal reservoir (Upper Permian; The Netherlands). In this case, the diagenetic study incorporates petrophysical and high-resolution X-Ray Fluorescence (XRF) core scanning data and is integrated in a solid sedimentological framework constrained by seismic, well-log and core records. The depositional texture was characterized at core and thin-section consisting on alternation of very fine- and very coarse-grained sand laminae, with the latter systematically cemented by anhydrite pore-filling cement. The identification of detrital anhydrite/gypsum grains pointed out to these grains as potential local source for anhydrite cement that precipitated in the coarser-grained, more permeable laminae. Due to the mechanical and chemical instability of anhydrite/gypsum detrital particles, their grain size may be used as an indicator of source distance, interpreted as nearby gypscretes. By combining these results with the prevailing west-southwest aeolian transport direction, a conceptual predictive model is proposed for distribution of these low-permeability streaks which reduce uncertainty in reservoir quality prediction. The multidisciplinary analytical workflow developed in this Thesis highlights the significance of diagenetic studies in reservoir quality prediction models. By using of outcrop-derived data, diagenetic analysis can be stressed to utmost when placed in a well-constrained depositional framework maximizing its use as input in upscaling reservoir properties.