Application of quantitative analysis of fluvial sedimentary architecture to improved facies and reservoir modelling workflows

Background. The development of a database approach to the quantitative study of fluvial sedimentary architecture enables structured queries to be made regarding the geometry, lithological heterogeneity and mutual relations of architectural elements that form fluvial successions (Colombera et al. 2012a). In applied contexts this information can be used to provide constraint to reservoir modelling work flows in a way that favours incorporation of geological realism in subsurface models as captured from analogue outcrop successions or modern depositional systems (Colombera et al. 2012b). Traditionally, it has been standard industry practice to run fluid-flow simulations based on models of subsurface geology in the absence of sufficient or appropriate primary data; in adopting such an approach, reservoir modellers often rely upon stochastically-derived simulations of architecture that are not necessarily founded on geological understanding. The recently developed approach of compiling a composite analogue using databasing principles addresses this shortcoming by providing data to establish and constrain stratigraphic relationships.
Project description. The primary aim, and major impact, of this research programme will be to use the established Fluvial Architecture Knowledge Transfer System (FAKTS – a large quantitative data database, developed in-house at Leeds, which describes aspects of fluvial sedimentary architecture) as the basis for devising novel and innovative, research-driven workflows for the development of a new generation of sophisticated facies and reservoir models for fluvial successions. This research project will offer reservoir modellers the option to replace stochastically-derived estimates of architectural variability with dimensions, geometries and lithologies of sedimentary bodies based upon real-world examples collected from outcrop analogues and modern systems, as described in the peer-reviewed literature. The method of metadata analysis that has been established in the Fluvial Research Group at Leeds (FRG) has two main benefits: 1) data are located all in one place and described to a common standard; and 2) co-collection in the database of information regarding the conditions affecting deposition allows developed models to utilize data only from a sub-set of samples that best correspond to the modelled situation. Thus, modellers will have the option to condition their models either using the standard stochastically-derived estimates of geological variability, or by using estimates derived from appropriate analogues. Further, the databasing approach will enable modellers to assess the ranges in model outputs against the boundary condition assumptions used to select the data, and will thus be able to establish which assumptions have the greatest effects. This will enable quantitative assessments of the relative significance of modelling assumptions, and so will enable better calibrated estimates of the risks associated with using these models to aid development decisions.