From outcrop to reservoir simulation model: Workflow and procedures

Subsurface reservoir models are typically limited by a lack of spatially accurate geometric data onbedform architecture and geometry, which are key controls on fluid flow. Outcrop analogues havelong been utilized as a source of such data, but the capture of sufficiently precise data has always beena challenge.This thesis demonstrates how advances in data capture and computer technology have made possiblethe collection of three-dimensional, high-resolution, digital geological data generated by ground basedlaser scanning (lidar) from steep and otherwise inaccessible outcrops. A complete workflow isdocumented — from outcrop selection through data collection, processing and building of virtualoutcrop models (VOMs) — to geological interpretation, data extraction and the building and testing ofgeocellular models using an industry-standard, reservoir-modelling software. Examples from analoguesystems such as the Roda Sandstone, Pyrenees, Spain, and Cretaceous deltaic systems from centralUtah, USA, are used to demonstrate the methodology.A parallel objective of the study has been to utilise the developed techniques to improve understandthe geometry, genesis and impact of shallow marine, deltaic clinothems. Clinothems are seawarddipping beds, formed when much of the sand in river-dominated deltas is laid down in distributarymouth bars that coalesce during the downstream accretion of the delta to form a delta front. Systematiccollection of data from VOMs has allowed for the compilation of a detailed database on clinothemsand associated clinoform geometry. Bed thickness and dip measurements have been collected fromtwo ancient, river-dominated deltaic depositional systems, namely the forced regressive PantherTongue and a high-stand portion of the Ferron Sandstone that both crop out in central Utah, USA.The extensive bed-thickness database of nearly 3000 measurements from 73 individual clinothems athundreds of meters to about 1000 m. scale has been used to quantify aspects of delta clinothem-beds.A decay gradient parameter permits the thinning of the clinothems to be described as a single numberand thus compared between beds, and measurements are used to calculate the average dip angle.Results demonstrate that clinothems in the Panther Tongue are much longer and more shallowlydipping than clinothems in the Ferron Sandstone. Both systems show similar patterns with regards togroups of gradient-values and groups of dip-values: Common boundaries exist between groups definedby gradient values and groups defined by dip values. A cyclic depositional pattern interpreted to berelated to autocyclic processes in deltaic mouth bars is documented, and three and nine stratigraphiccycles or bedsets have been identified in the studied parts of the Ferron Sandstone and the PantherTongue, respectively. The differences between the two systems are attributed to their sequencestratigraphic positions and hence accommodation.Deterministic geocellular reservoir models have been built from surfaces interpreted in the VOMs.The models have been used to test the influence of dipping siltstones heterogeneities on simulatedreservoir performance in a reservoir modelling software. Siltstone drapes on clinothem sandstone bedshave been modelled and various populations of holes have been placed in the siltstone beds (0, 30, 60,90 and 100%). The effects of siltstone permeability (1, 0.1, 0.01 and 0.001 mD) have also been tested.A total of 41 individual models were built and flow simulated. Results quantify how the portion ofholes governs the production rate/ recovery factor in the Panther Tongue models. Permeability valuesare more important in the Ferron models, although they are still influenced by the number of holes.Steeper dipping and closer spaced clinothems of the highstand system tract lower the recovery factorby several tens of percent if the related heterogeneities are all or close to continuous and have lowenough permeability