MULTI-SCALE ANALYSIS OF SIMULATED MEANDERING FLUVIAL SYSTEMS: IMPLICATIONS OF DEPOSITIONAL CONDITIONS ON 3D GEOLOGICAL MODEL DEVELOPMENT

Building 3D geological models is a key step in managing groundwater, geothermal, and hydrocarbon resources, and facilitating carbon sequestration. Strategies for populating geological models draw on observations from core to basin scale. The finer the scale of observation, the higher the potential resolution of a geological model. However, cost of data collection, model simulation and computation increases with the resolution of 3D geological model. Here we use numerical modelling to investigate the trade-off between increased model resolution and simulation accuracy in meandering fluvial deposits. We explore how depositional conditions can be considered to help constrain the appropriate observation scale for building an optimal 3D geological model. Using FLUMY software, we simulated 270 3D models of meandering fluvial depositional successions with varying coarse-to-fine sediment inputs and river scales. Keeping the grid lag [10 m x 10 m x 0.1 m] and model domain [2500 m x 2000 m x 50 m] same for each simulation, we populated the models with stochastic rock properties applied at a facies scale (i.e. different distributions of properties for channel and overbank deposits) and also with an ordered sequence of rock properties for each facies derived from a compiled database of sedimentary logs from published literature. Stochastic and ordered models were compared using static connectivity and fluid-flow based dynamic heterogeneity parameters.

Results show that when basins were filled by small channels and/or high coarse sediment inputs, performance differences between the stochastic and ordered models are negligible, indicating that simpler, stochastic assignment of rock properties is sufficient. However, in some cases (large channels and fine sediment), the difference between stochastic and ordered rock properties is significant. This work demonstrates how properly scaling sedimentary deposits can guide 3D geological model development and shows how depositional conditions can influence subsurface heterogeneity. Thus, geological conditions known a-priori can be used to determine the most appropriate assumptions for optimal 3D geological model development. The modelling approach presented here can be used to optimize sampling resources in subsurface applications, such as carbon sequestration.