Characterization and 3D reservoir modelling of fluvial sandstones of the Williams Fork Formation, Rulison Field, Piceance Basin, Colorado, USA

This study describes the stratigraphic characteristics and distribution of fluvial deposits of the Upper Cretaceous Williams Fork Formation in a portion of Rulison Field and addresses 3D geologic modelling of reservoir sand bodies and their associated connectivity. Fluvial deposits include isolated and stacked point-bar deposits, crevasse splays and overbank (floodplain) mudrock. Within the Williams Fork Formation, the distribution and connectivity of fluvial sandstones significantly impact reservoir productivity and ultimate recovery. The reservoir sandstones are primarily fluvial point-bar deposits interbedded with shales and coals. Because of the lenticular geometry and limited lateral extent of the reservoir sandstones (common apparent widths of similar to 500 - 1000 ft; similar to 150 - 300 m), relatively high well densities (e. g. 10 acre (660 ft; 200 m) spacing) are often required to deplete the reservoir. Heterogeneity of these fluvial deposits includes larger scale stratigraphic variability associated with vertical stacking patterns and structural heterogeneities associated with faults that exhibit lateral and reverse offsets. The discontinuous character of the fluvial sandstones and lack of distinct marker beds in the middle and upper parts of the Williams Fork Formation make correlation between wells tenuous, even at a 10 acre well spacing. Some intervals of thicker and amalgamated sandstones within the middle and upper Williams Fork Formation can be correlated across greater distances. To aid correlation and for 3D reservoir modelling, vertical lithology proportion curves were used to estimate stratigraphic trends and define the stratigraphic zonation within the reservoir interval. Object-based and indicator-based modelling methods have been applied to the same data and results from the models were compared. Results from the 3D modelling indicate that sandstone connectivity increases with net-to-gross ratio and, at lower net-to-gross ratios (< 30%), differences exist in the cumulative volume of connected sandstone bodies between the indicator- and object-based lithology models. Therefore, the types of lithology-modelling methods used for lower net-to-gross ratio reservoir intervals are important.