Groundwater Modeling in Alpine Karst Aquifers: An Ensemble Approach
AuthorFandel, Chloe Alexandra
AdvisorFerre, Ty P.A.
MetadataShow full item record
PublisherThe University of Arizona.
RightsCopyright © is held by the author. Digital access to this material is made possible by the University Libraries, University of Arizona. Further transmission, reproduction, presentation (such as public display or performance) of protected items is prohibited except with permission of the author.
AbstractOne of the greatest challenges in hydrogeology is the lack of detailed information about the spatial configuration of the subsurface. In karst systems especially, groundwater flow is strongly controlled by difficult-to-map conduit networks. Spatially discretized information about the subsurface is also a key input for numerical models, which are increasingly widely used to guide decision-making. Although many researchers agree that the spatial structure of an aquifer is a major source of model prediction uncertainty, in practice, water resource managers often must make decisions based on models built from limited information, with minimal uncertainty analysis. This dissertation presents a new approach to modeling karst, in which the uncertainties inherent in working with incompletely-mapped subsurface drainage networks are explicitly recognized, and in which useful results can be obtained despite limited data availability. Methodological advances to make this approach possible are introduced, including: 1) pyKasso, an opensource, computationally-efficient conduit network model with minimal data requirements, capable of generating many plausible network maps of the same system using anisotropic fast marching methods; and 2) a comprehensive ensemble generation workflow, linking a geologic modeling step, a conduit modeling step, and a pipe-flow modeling step, and yielding a diverse ensemble of spring discharge predictions for the same system. These tools were then applied to address several questions in a complex alpine karst catchment: the Gottesacker long-term study system (Germany/Austria). Findings include: 1) confirmation that structural uncertainty has a larger influence than parameter uncertainty on prediction uncertainty for this system; 2) identification of the most informative locations for further field data collection to discriminate between competing models of the network structure; and 3) support for the hypothesis that the lowest-elevation spring in the system (Sägebach) is also the youngest, and that the inactive portions of the mapped Hölloch cave system developed before the spring came into existence.
Degree ProgramGraduate College