The heterogeneity of fluvial systems is a key parameter in sedimentology due to the associated impacts on flow performance. In a broader context, fluvial reservoirs are now targets for CO2 storage projects in several sedimentary basins (Paris Basin, North German Basin), thus calling for detailed characterization of reservoir behaviour and capacity. Fluvial reservoirs are a complex layout of highly heterogeneous sedimentary bodies with varying connectivity, depending on the sedimentary history of the system. Reservoir characterization must determine (a) the nature and dimension of the sedimentary bodies, and (b) the connectivity drivers and their evolution throughout the stratigraphic succession. Based on reservoir characterization, geological modelling must account for this information and can be used as a predictive tool for capacity estimation. Flow simulation, however, describes the reservoir behaviour with respect to CO2 injection. The present work focuses on fluvial reservoir performance and was carried out as part of a PhD (2008-2011) dedicated to the impact of sedimentary heterogeneity on CO2 storage performance. The work comprises three steps: ● Reservoir characterization based on detailed fieldwork (sedimentology and sequence stratigraphy) carried out in Central Arabia on the Minjur Sandstone. Twelve depositional environments and their associated heterogeneity are identified, and their layout is presented in a high-resolution sequence stratigraphy analysis. This step is summed up in a 3D geological model. ● Conceptual modelling based on this field data, using gOcad software and an in-house python code. The purpose was to study, for a given architecture, the impact of sedimentary heterogeneity on storage capacity estimations using two models: one with heterogeneity within the sedimentary fill (model A); the other without heterogeneity within the sedimentary fill (model B). A workflow was designed to estimate and compare the storage capacities for a series of some 50 scenarios. The results show that a strong compartmentalization, due to a shaly barrier, may decrease storage capacity by 11 to 25 percent. ● Flow-simulation of an 8-scenario sample extracted from the 50 possible scenarios. In contrast to the static modelling estimated capacities, the preliminary flow-simulation results indicate that capacity remains similar whichever model is applied (A or B). This is because the scale of the heterogeneity is similar to the extent of the CO2 plume, meaning that heterogeneity does not affect the amount of injected CO2 that can be stored in the sedimentary body. Nevertheless, connectivity strongly influences storage capacity, as determined by the 8 scenarios (model A) in which the total amount of CO2 injected ranges between 7 and 12 Mt over a 50-year period. Moreover, heterogeneity significantly increases pressure build-up, and may strongly disrupt the hydrodynamics in the aquifer.