openKARST: A novel computational modeling tool for flow and transport in complex karst conduit networks

We present a newly developed flow and transport simulator for karst systems, designed to model complex flow dynamics and tracer transport in large-scale networks. The simulator is based on the Saint-Venant equations and integrates advanced hydraulic modeling and particle tracking algorithms to study the interplay between the physical properties of karst conduits, their large-scale network structure, and the resulting flow and transport behavior. The simulator accommodates both steady-state and transient flow scenarios under free-surface and pressurized conditions. Turbulent flows are modeled using the Darcy-Weisbach equation, supported by classical friction models such as the Churchill and Colebrook-White equations. Validation against a broad range of analytical solutions and flow dynamics in one of the largest cave systems in the world, the Ox Bel Ha cave, confirms the robustness of the approach.

Using field data obtained from 3D lidar scans of cave systems, we extract geometries to build high-resolution network models and investigate how resolution impacts flooding signals. Specifically, we analyze how downscaling to lower resolutions, resulting in fewer conduits with averaged properties such as diameter, hydraulic radius, and roughness, alters critical features like bottlenecks and their influence on flow propagation. Bottlenecks, which play a significant role in controlling flow rates and local hydraulic gradients, can disappear or be muted when the geometric complexity of the network is reduced. This smoothing process effectively reduces the spatial variability in conduit dimensions and frictional resistance, leading to changes in the timing, magnitude, and spatial distribution of flooding signals. The disappearance of bottlenecks at lower resolutions may result in a more homogenized flow regime, potentially masking the true hydrodynamic behavior of the original network. Understanding these impacts is critical for accurately modeling flow dynamics in karst systems and assessing the trade-offs between computational efficiency and the complexity of hydrological predictions.