Search

Submarine fan deposits are volumetrically the largest sediment accumulations on Earth and host significant hydrocarbon reserves. Extensive research has documented the bed‐scale architecture of high sand‐to‐mud ratio, proximal and axial environments, which can have bed thicknesses of several metres; however, less well‐understood are thin‐bedded turbidites, which are typically lower N:G and deposited in more distal environments. Conceptual models assume that lobe‐fringe‐to‐basin plain environments consist of tabular, sheet‐like beds that extend out continuously and predictably over long distances—up to several kilometres. Extensive lateral continuity, however, is not necessarily reflected in ancient outcrop analogues. This study seeks to apply a quantitative approach to the characterisation of thin‐bedded turbidites to assess the impact of multi‐scale heterogeneity on reservoir predictability. The sea‐cliff outcrop exposures of the Upper Cretaceous Point Loma Formation in San Diego, California, exhibit a wide range of bed thicknesses and stratigraphic architecture, which have been used to interpret an off‐axis‐to‐fringe depositional environment. The study area spans 700 m of laterally continuous outcrop, across which 10 correlated stratigraphic sections are used to quantify changes in metrics such as bed thickness, N:G, lithofacies proportions, etc. Results of this study demonstrate that thin sand beds experience both lateral facies changes and rapid thickness changes more frequently than conceptual models would predict. A single measure of lateral heterogeneity does not reflect the true architecture of sandstone beds, and significant information is lost when beds are correlated over ten to hundreds of metres. Sands are commonly deposited in irregular, 'finger‐like', planform geometries, which compounds at the lobe element scale and influences lateral lithofacies predictability. This study of the Point Loma Formation offers high‐resolution bed‐to‐element scale data, which may be used as inputs for reservoir models and horizontal facies predictions in both conventional and unconventional hydrocarbon reservoirs. [ABSTRACT FROM AUTHOR]

Current knowledge of submarine fan deposits has historically relied heavily on qualitative field and subsurface observations and interpretations, but recent studies using statistical analyses have enhanced the understanding of submarine fan sub‐environments, including the degree of confinement, stratigraphic patterns and potential control factors. The purpose of this study is to improve the quantitative understanding of synorogenic submarine fan deposition at foreland basin settings with a statistical approach. A suite of statistical methods is integrated and developed (Hurst Statistics/rescaled range analysis, bed thickness frequency distribution analysis, Markov Chains and time‐series analysis), and applied to the well‐understood Pennsylvanian lower Atoka submarine fan system in the Ouachita Mountains, United States, for this purpose. The results of the Hurst Statistics and bed thickness analyses corroborate qualitative interpretations that: (i) the lower Atoka is lobe‐dominated; and (ii) the south‐eastern (wedge top) portion of the system is more strongly confined than the northern (foredeep) portion. The Markov Chains and time‐series analyses reveal the prevalence (56% of the measured sections) of stratigraphic orderliness and cyclicity; these results are used to discuss potential control signals, which are otherwise difficult to distinguish, on the turbidite sandstone recurrence cycles. The results of this study demonstrate that these integrated statistical methods can be utilized to quantify depositional interpretations and extract hidden information using outcrops with limited exposures. [ABSTRACT FROM AUTHOR]

The partitioning of different grain‐size classes in gravity flow deposits is one of the key characteristics used to infer depositional processes. Turbidites have relatively clean sandstones with most of their clay deposited as part of a mudstone cap or as a distal mudstone layer, whereas sand‐bearing debrites commonly comprise mixtures of sand grains and interstitial clay; hybrid event beds develop alternations of clean and dirty (clay‐rich) sandstones in varying proportions. Analysis of co‐genetic mudstone caps in terms of thickness and composition is a novel approach that can provide new insight into gravity flow depositional processes. Bed thickness data from the ponded Castagnola system show that turbidites contain more clay overall than do hybrid event beds. The Castagnola system is characterized by deposits of two very different petrographic types. Thanks to this duality, analyses of sandstone and mudstone composition allow inference of which proportion of the clay in each of the deposit types was acquired en route. In combination with standard sedimentological observations the new data allow insight into the likely characteristics of their parent flows. Clean turbidites were deposited by lower concentration, long duration, erosive, muddy turbidity currents which were more efficient at fractionating clay particles away from their basal layer. Hybrid event beds were deposited by shorter duration, higher‐concentration, less‐erosive sandier flows which were less efficient at clay fractionation. The results are consistent with data from other turbidite systems (for example, Marnoso‐arenacea). The approach represents a new method to infer the controls on the degree of clay partitioning in gravity flow deposits. [ABSTRACT FROM AUTHOR]

Recent studies of turbidite bed thickness distributions have demonstrated power-law as well as log-normal statistical distributions. The different distributions may reflect different fan processes and environments and, therefore, could be used as a quantitative method to help identify those environments, including those devoid of sequential patterns. The cumulative distributions of well-known turbidite deposits spanning a range of interpreted fan subenvironments are used to illustrate the potential correlation between cumulative distribution and environments. Assuming that power-law distributions may, for some systems, be the primary input signal, one-dimensional modelling allows semi-quantitative characterization of the effects of different fan processes such as erosion and bed amalgamation. Environments indicative of different fan processes may be characterized based on the degree to which processes have acted as a ‘filter’ to modify the assumed power-law distribution systematically. This model of the effect of fan processes on the power-law distribution is used to help to account for bed thickness distributions observed in several field sites. [ABSTRACT FROM AUTHOR]