TOWARD MORE REALISTIC REPRESENTATION OF DISCRETE CONDUIT FEATURES IN DISCRETE-CONTINUUM FLOW AND TRANSPORT MODELS OF KARST SYSTEMS (Invited Presentation)

Discrete-continuum models (DCMs) are considered as suitable distributed numerical tools for process-based simulation of karst systems. Therefore, inverse application of such models has been served for proof of concepts and system characterization in many karst and pseudokarst systems all over the world. DCM results were found promising for many regional and large-scale applications. However, effects of unaccounted processes and system features potentially have been overlooked. Recent joint inversions of groundwater flow, heat, and mass transport data for Freiheit Spring, a small-scale karst system in Minnesota, US, revealed that the effect of unaccounted model characteristics (e.g., noncircular conduits) can be substantial. While a statistically robust result could be achieved for such a spatiotemporally small-scale investigation, it was demonstrated that some model parameters are compensating for the unaccounted processes and/or features of the system.

Conduit cross-sectional shapes (CCSs) are presumed as circular in DCMs, though speleological investigations suggest that this assumption may not hold true in many cases, especially since the shape of rounded cave passages may yet be affected by sediment fillings. The focus of this research is to investigate the importance of CCSs on the flow, heat, and solute transport signals in karst systems.

Firstly, three CCSs, including circular, elliptical, and round-ended flat cases were assumed and compared in terms of hydraulic diameters for full- and partially-filled-pipe systems, as well as discharges at different imposed hydraulic gradients and conduit volumes. Results of such simple sole pipe flow systems provide understanding of the impact of CCSs.

Next, as a part of ongoing research, the MODFLOW Conduit Flow Process Version 2 (known as CFPv2) is modified to account for the aforementioned CCSs. Test simulations of flow and transport signals are able to demonstrate the effect of the missed system feature on the observed system flow and transport behavior, likely masked by the process length- and time-scales, as well as tributary mixings at large scales. We are of the opinion that the most probable CCSs should be investigated, based on the geological and speleological information, in order to build a congruent DCM, especially at small scale sites.