In the context of field appraisal in turbidite environment, with the excessive costs associated to deepwater drilling, well data is limited. High-quality 3D seismic data is a key driver for reservoir characterisation and modelling. Seismic volume attributes can be used to attempt a quantitative prediction of facies heterogeneities within the reservoirs. The present work develops a method for enhanced conditioning of geostatistical facies modelling to seismic attributes, thus combining techniques for a better representation of reservoir heterogeneities.
This methodology was developed on a deepwater turbidite field in conceptual phase of development. Only one well was drilled, and a seismic attribute cube of pseudo-Vclay (pseudo-volume of clay) is available. The main reservoir unit consists in a wide fairway containing erosive-constructive unitary stacked channels.
In classical approaches, sand proportions in the reservoir are derived from seismic attributes by calibration using well data. The idea behind the new method presented here is to calibrate the pseudo-volume of clay attribute to actual sand proportions using a training image, getting around the lack of extensive well data for calibration. The training image provides a conceptual target derived from sedimentological models for the statistical distribution of sand proportions in the reservoir. The outcome is a reservoir property of calibrated sand proportions that can be used as soft auxiliary data for facies modelling using multiple-point statistics.
Seven scenarios were designed to account for geological uncertainties, and multiple stochastic realizations were generated for each scenario. Connected oil volume inferred from well testing was used as the main controlling parameter for evaluating the resulting facies models. Scenarios considering high channel stacking and large channels came best at honouring dynamic data. Other scenarios with models presenting a connected volume too low in the fairway, if considered or proven realistic, puts into question the role of levees in the dynamic behaviour of the reservoir.
In the context of field appraisal in deepwater turbidite environment, with the excessive costs associated to deepwater drilling, well data is limited. 3D seismic data of very good to excellent quality is a key driver for reservoir characterisation and picking geological bodies and architectures. These architectures are essential elements that in turn drive the facies and petrophysical modelling workflows.