Field Scale Automated Calibration and Sensitivity Analysis of Forward Stratigraphic Models

This study discusses the combination of forward stratigraphic modelling, experimental design and surface modelling techniques in order to produce multiple plausible geological models calibrated at wells. Sensitivity analysis conducted on influencing environmental parameters allows a better understanding of the lateral heterogeneity of carbonate reservoirs using process-based approaches. It also provides a better control on future dynamic reservoir simulations. The deterministic 4D forward models incorporate bathymetric and thickness data in order to calculate basin subsidence through time. Carbonate production laws versus time, depth and wave/drift energy allow generating complex carbonate geometries as mounds, reefs, clinoforms. Such geometries are difficult to model with stochastic approaches that do not account for geological processes influencing carbonate production and transport. This innovative workflow is divided in three main steps: 1. A reference case forward stratigraphic model is manually calibrated to well data (thickness and texture calibration); 2. Experimental design and surface modelling approaches are used to screen all the inputted uncertain environmental parameters in order to assess the most influencing ones; 3. A second run of automated simulations using the previous techniques are used to generate highly calibrated and geologically plausible models, later assessed through sensitivity map analysis. The application of this workflow as alternative/complement to stochastic methodologies increases the confidence and predictability of static and dynamic facies models underlining the non-uniqueness of geological models.