Towards simulating solute transport in complex, regional-scale fracture networks: a rapid upscaled approach

Currently, the most common approaches for simulating solute transport in fractured aquifers are the single- or dual-continuum and the discrete fracture network (DFN) methods. However, continuum approaches often lack accuracy due to averaging, whereas DFN approaches may be computationally prohibitive for large-scale fracture networks. To address these challenges, this study presents an Upscaled Fracture Network (UFN) model, developed by discretizing complex fracture networks into elementary volumes, identifying solute transport flow channels and calculating breakthrough curves within an elementary volume. The identified flow channels within the micro-scale DFN are collectively employed to construct the residence time at the macro-scale DFN. Validated against a random walk particle tracking (RWPT) DFN-based approach, the UFN model accurately captures solute transport processes in saturated fracture networks at the macro scale, and represents a significant advancement in simulating solute transport in complex, regional-scale aquifers due to its computational efficiency, simple implementation, and high level of accuracy. • An Upscaled Fracture Network (UFN) is developed to model solute transport at the macro-scale fracture networks by obtaining the solute transport migration in the micro-scale DFN and translating it to the macro-scale DFN. • The UFN model can accurately capture solute transport processes occurring at the macro scale in saturated fracture networks, and potentially more accurate and computationally efficient, approach to upscaling continuum transport models. • The UFN model represents a significant advancement in simulating solute transport in complex, regional aquifers due to its computational efficiency, simple implementation, and high level of accuracy. • The UFN model arrived at a solution in approximately 25% of the time required for the RWPT model with 5,000 particles released. The effect of this efficiency becomes even more important when more particles are released, and in larger, more complex networks. • The UFN model is capable of simulating solute transport in DFNs with a significant variation in the statistical properties across the network. Such cases can be effectively managed in the UFN model by dividing the macro-scale region into subregions where network properties remain approximately constant within each subregion.