Reducing predictive uncertainty in log-derived water saturation models in a giant tight shaly sandstones – A case study from Mesaverde tight gas reservoir

• The Indonesian model was found as the best traditional model to predict water saturation in tight shaly sandstones.
• The proposed calibration coefficient has the capability of calibrating all traditional models with tight shaly sandstones.
• The introduced calibration coefficient makes the Indonesian as the best model that can reduce the saturation uncertainty.

A log-derived calibration coefficient is introduced to calibrate conventional water saturation models in tight shaly sandstone reservoirs in the Rocky Mountains, western United States, in this study.It is implemented for calibration of eight traditional saturation models in this giant gas field to prove its validity. The application of this calibration coefficient indicates that it is a simple and strong tool for improving of water saturation estimation in tight shaly sandstones, particularly for zones with high GR and/or low true resistivity. The proposed calibration coefficient is based upon an expression which depends on true resistivity log, neutron and density logs to take total porosity and gamma ray log. It has a significant advantage i.e. independency from electrical rock properties (Archie's parameters): m (cementation factor), n (saturation exponent) and a (tortuosity factor). It means that the variety of them cannot affect the operation of the introduced calibration coefficient.In this study, the calibration results of the eight traditional water saturation models are compared with core results for 2579 data points that are taken from fourteen wells (39 different zones) from Mesaverde in five western USA basins. This study shows that the new approach that is introduced is more efficient as it reduces the uncertainty associated with water saturation estimation in zones with high GR and/or low true resistivity than zones with low GR and/or high true resistivity in tight shaly sandstones.