Dynamic capillary barrier failure in stratified media: A dimensionless framework for LNAPL redistribution under water...

Subsurface contamination by light non-aqueous phase liquids (LNAPLs) poses long-term environmental risks, which is significantly influenced by the dynamic coupling of transient water table fluctuations (WTFs) and stratigraphic heterogeneity. To assess aquifer vulnerability and decouple these multiscale physical processes, we propose a macroscopic dimensionless framework governing dynamic LNAPL redistribution and capillary barrier failure. Within this framework, three key dimensionless parameters are established: a maximum spatial coupling index (ƞ_max) defines the spatial continuity prerequisite for transient pressure transmission; a modified Bond number (Bo*) determines the macroscopic threshold for hydraulically driven interfacial capillary barrier failure; and a macroscopic capillary number (Ca_macro) scales the kinetic competition between vertical contaminant penetration and lateral spreading. By factoring out the constant geometric constraints of our specific quasi-2D experimental setup, the dynamic component of this kinetic regulator is further isolated as a system-level parameter (Ca*). To validate this theoretical approach, systematic 2D sandbox experiments were conducted within stratified systems featuring distinct coarse-over-fine (C-F) and fine-over-coarse (F-C) interfaces under varying WTF amplitudes, initial water tables, and fluctuation rates. The experimental results reveal that LNAPL redistribution in F-C structures follows a threshold-controlled mechanism governed by macroscopic displacement pressure, closely matching the dual critical breakthrough conditions of ƞ_max ≥ 1.0 and Bo* > 1.0. Conversely, C-F structures lacking interfacial capillary barriers exhibit a rate-dependent allocation mechanism, where lower Ca* values significantly promote lateral hazardous plume expansion via pore-scale capillary wicking compared to high-rate conditions. External validation across multiple fluid datasets confirms that this framework provides a first-order predictive boundary, offering essential guidelines for evaluating contaminant fate, capillary barrier stability, and overall environmental risk in highly dynamic groundwater systems.