The Crystal Viking Field, which was discovered in 1978, is a compound incised valley-fill, comprising two pools. Reservoir characterization, geostatistical modelling and simulation were undertaken to better understand the reservoir, and to identify new business opportunities.

The reservoir interval within the valley-fill complex comprises six major lithofacies: conglomerate; conglomeratic sandstone; medium-grained sandstone; fine-grained, laminated sandstone; fine-grained, shaley sandstone; and shale. Each of the lithofacies corresponds to a different sub-environment within the valley-fill. Re-examination of existing core descriptions (113 cored wells out of 148 wells in the field), and integration of the core descriptions with numerous cross-sections and slice maps, and the use of vertical proportionality curves, resulted in a different interpretation of the internal architecture, and number of fills within the valley-fill relative to previously published interpretations. The interpretation here is one of a tidal-fluvial dominated estuarine fill with a marginal mudflat.

Upon examination of the individual lithofacies and their possible interpretation (e.g., 3, 4, 5 or more individual fills), it was concluded that there was no reliable way to determine correct interwell correlations. Additionally, no permeability barriers were identified between lithofacies using probe permeametry results. As a result, instead of trying to identify and correlate different ‘arbitrary fills' within the ‘A pool', we decided to constrain the model by only the top and bottom surfaces of the ‘A pool', and allowed the geostatistical model to fill space, controlled only by well data.

Each of the major lithofacies forms a fundamental geological element of the reservoir model, each with its own unique porosity and permeability distribution. Porosity and permeability were derived from log analysis, calibration and modelling using core analysis and probe permeameter data. From this information, a non-stationary geological model was built iteratively for ‘A' and ‘H' pools individually until a satisfactory geostatistical model was created.

A reservoir simulation model was derived from the up-scaled geostatistical model. A global history match was achieved within a few weeks, suggesting that we had done a good job with the geological modeling. Previous attempts at modelling the Crystal Field using a deterministic approach were less successful.

As a result of using reservoir characterization , geostatistical modeling and reservoir simulation, new opportunities for infill drilling and recompletions/workovers have been identified, and there is now a different understanding about the remaining potential in the field and what it will take to optimize the waterflood in the Crystal Field.

AAPG Search and Discovery Article #90937©1998 AAPG Annual Convention and Exhibition, Salt Lake City, Utah