Fluvial sandstones are important hydrocarbon reservoirs in the Middle East and world-wide. However the significant lithological heterogeneities and lateral continuity are main limitations when considering reservoir development plans. Furthermore low net-to-gross sequences are usually beyond seismic resolution as only the stacked sandstones of the incised valley fills provide sufficient acoustic contrast to be recognised seismically. Major problems are usually encountered in defining sand volume, connectivity and recoverable hydrocarbons. All these factors are critical when considering field management plans and are related directly with the definition of complex spatial configurations, constrained to geological information usually via a genetic classification scheme.
When insufficient information is available the optimisation of modelling procedures becomes of utmost importance. In those cases the integration of all the available information with stochastic modelling techniques minimises uncertainties in the input values used as reservoir parameters for building predictive models for the distribution of sandbodies, flow properties and recoverable hydrocarbons.
This study concerns the development of an integrated methodology for modelling fluvial sequences presenting low net-to-gross ratios, in which channel belts are enclosed within flood plain fines. The main steps are:
characterisation of channel main morphological features from well data and outcrop analysis;
stochastic simulation of channels geometry conditioned to local anisotropy of meander orientations (dipmeter measures, FMI/FMS, etc.);
pos-processing of the stochastic images of the meander-belt structures using a morphological filter to clean noisy features.
An example from the Middle East is presented in order to illustrate the application of the methodology to a typical fluvial environment presenting complex characteristics.
Fluvial channel sandstones constitute important oil and gas reservoirs in several hydrocarbon provinces around the world. Most fluvial channel deposits occur in laterally confined elongate channel belts as a consequence of the distributary character of the fluvial system. There is, however, considerable variation in modern fluvial channel morphology and consequently in the geometry of associated channel belt deposits.
Quantified information on the geometry of sandstone bodies, their internal heterogeneities and relative positioning is a necessary prerequisite for the realistic modelling of a fluvial reservoirs. The general effort towards improved reservoir description and simulation includes not only the development of new mathematical and computational tools but also widespread studies of analogue outcrops and compilation of geological databases containing information concerning alluvial architectures and sedimentary bodies.
Collection of large data sets which can provide information to support geomathematical modelling of reservoir sequences is therefore important and necessary. However, the outcrop related data which is going to be used as input should fulfill the requirements for the modelling techniques and be processed accordingly. Usually two types of information is required: information concerning the geological setting and information related to the geological framework and sedimentary architecture of the deposits at different scales.