Workflow using sparse vintage data for building a first geological and reservoir model for CO₂ geological storage in deep saline aquifer. A case study in the St. Lawrence Platform, Canada

Among all geological CO 2 storage possibilities, deep saline aquifers are of great interest due to their worldwide repartition and their important storage volume. We present a workflow using available vintage data with poor 2D seismic coverage for building a first geological and reservoir model for CO 2 geological storage in the deep saline aquifers of the St. Lawrence Platform in the Bécancour area (Québec, Canada). In order to optimize the sparse available geoinformation using a geostatistical method, we krige the tops of the geological formations recorded at 11 wells using surfaces modeled from seismic horizons picked on 99.4 line‐km of 2D seismic reflection data. Modeled geological horizons show a good compromise between the geometric structure expressed by the variograms and the interpreted variations evaluated from seismic horizons. Using available well logs, distribution of porosity and permeability are computed for generating multiple realizations of the petrophysical properties of the targeted aquifer by sequential Gaussian simulations. The scarcity of available petrophysical data in the targeted aquifer generates high variability between the different realizations. Due to this uncertainty, the population of the 3D geological model with petrophysical properties that are required for further geostatistical simulations of CO 2 injection do not allow to achieve reliable results. The methodology presented in this paper shows the possibilities and limits of using vintage data, and provides evidence that geophysical data acquired in a 3D fashion are important to fully characterize a reservoir for CO 2 geological storage.