Estimating the geometry and sedimentary heterogeneities of carbonate platforms is a very challenging task. In recent decades, numerical stratigraphic forward models have been developed to provide a quantitative view of sedimentary processes with the resulting stratigraphic architecture. Although used in many academic and industrial studies, this numerical approach is rarely used in everyday exploration and appraisal workflows. The calibration of these models on available seismic and well data is indeed a challenging and time-consuming process.

Here we use the DionisosFlow model to simulate a carbonate platform typical of the Miocene formations observed in Central Luconia, South China Sea. We build a reference simulation or virtual ground truth from literature data, using for instance the Haq et al. (1988), Laskar et al. (2011), and Bosscher and Schlager (1992) curves, which prove indications on sea-level variations, orbital parameters, and carbonate production as a function of water depth, respectively. We analyze the sensitivity of the simulation results to the parameters controlling the physical processes: accommodation space creation, carbonate production, and sediment transport.

Results show that stratigraphic modeling makes it possible to test the concepts of sequence stratigraphy in a virtual digital world, thus opening the possibility of testing the sensitivity of the different characteristics of a sedimentary system to physical parameters. We also propose an innovative approach to using this stratigraphic modeling in operational cases. The first step is to identify a diachronous geological body such as the carbonate platform sensu stricto, which is easily identifiable using seismic data. A comparison of the geometry of this geological body with the thickness maps derived from the seismic interpretation provides a first regional metric controlling the shape of the studied sedimentary systems. The second step is to validate the simulation results with well data, and, in particular, facies from the log interpretation. The computation of vertical proportion curves (VPC) in the vicinity of wells facilitates understanding the local variability of facies in the digital world as well as in the real world. This second metric, based on a comparison of VPCs, makes it possible to assess sequences and sedimentary heterogeneities and to define much smoother cost functions, thus facilitating the use of automatic optimization algorithms.

In conclusion, this work shows that numerical stratigraphic forward modeling is a tool that reproduces stratigraphic concepts using simple physical laws. Through the use of well and seismic metrics adapted to the resolution of the model, use of this numerical approach in daily exploration work to complement the interpretation of seismic and well data provides a quantitative stratigraphic view of the studied area to better access uncertainties and risk in exploration.