Dual-domain modeling of infiltration dynamics in the vadose zone of karst systems using film-flow theory – Investigation of compartment-specific parameter sensitivities and film-flow parameters

Karst aquifers provide approximately 9.2 % of the global population with drinking water (Stevanović, 2019), and therefore, their proper water resource management is an important task. The assessment of available water resources requires both, the identification of the structure of the aquifer geometry including phreatic and vadose zone and the characterization of hydraulic parameters. However, the heterogeneity of karst systems poses a challenge for the accurate characterization based on numerical modeling approaches and obstructs risk assessment efforts. Highly conductive features within the vadose zone offer a domain for rapid infiltration via preferential pathways that may not be accurately recovered by classical instruments, e.g., the Richards equation, since gravity-driven flow regimes may prevail. Hence, we propose a distributed dual-domain modeling approach that accounts for both the diffuse infiltration through a porous matrix and film-flow on fracture surfaces. Phreatic flow, including flow within a porous matrix, flow in conduits, and an exchange between these domains, is realized by the numerical modeling framework implemented in MODFLOW-CFPv2. Furthermore, this study presents compartment-specific parameter sensitivities during infiltration events in a synthetic karst model and methodology to determine film-flow parameters based on field data, i.e., precipitation time series and water table fluctuations. The global sensitivity analysis highlights the influence of film-flow parameters, i.e., the limiting fracture facial area along the z-axis, F_lim, and an activation threshold, q_thr, while a recharge pulse persists. The delay between the commencement of infiltration and the hydraulic response at the water table, t_lag, may be site-specific and relies on the availability of observation data with proper temporal and spatial resolution.