Evaluating Subsurface Parameterization to Simulate Hyporheic Exchange: The Steinlach River Test Site.

Accurate simulation of hyporheic exchange remains a challenge due to subsurface heterogeneity. Prior research has shown that groundwater and river water interactions are influenced by subsurface properties. However, the impact of different subsurface parameterization strategies on model accuracy is unclear. This gap motivated a closer examination of how structural choices affect predictive errors. No prior work had resolved the trade-offs between intrinsic and epistemic errors in this context. Existing models often assume homogeneity, which may limit accuracy. That uncertainty drove the need to test various subsurface representations. It was already known that heterogeneity affects flow paths and residence times. This study builds on that foundation to explore modeling trade-offs.

The purpose of this study was to evaluate how subsurface parameterization choices affect hyporheic exchange simulations. The Steinlach River Test Site in Germany was selected as a case study. The researchers aimed to compare homogeneous and complex subsurface structures. They wanted to assess how these choices influence predictive errors and uncertainties. The study focused on transit time distributions as a key metric. The goal was to determine the most practical parameterization approach. This approach should balance model accuracy with data availability. The study also aimed to identify which parameterization methods reduce uncertainty most effectively.

The study used a fully coupled surface water-groundwater model to simulate hyporheic exchange. The Steinlach River site was modeled using different subsurface parameter structures. A highly detailed model served as a reference or 'virtual reality'. The researchers compared homogeneous and zonated subsurface representations. They also tested interpolated heterogeneous parameter fields and multiple-point geostatistics. Each model's performance was evaluated based on transit time distributions. Predictive errors and uncertainties were quantified for each structure. The comparison focused on how well each model reproduced the reference data.

The study found that overly simple parameter structures led to large intrinsic errors. Increasing subsurface complexity reduced intrinsic errors but introduced epistemic ones. Zonated models reproduced transit time distributions well when geological structures were known. Interpolated heterogeneous fields provided the best balance between error types. Multiple-point geostatistics produced the most uncertain results. However, these uncertainties could be reduced with additional data. Flux measurements were identified as particularly useful for improving accuracy. The findings suggest that model complexity should match available data and objectives.

The authors concluded that subsurface parameterization choices significantly affect hyporheic exchange simulations. They found that intrinsic errors are unavoidable with overly simple models. Epistemic errors can be reduced through better data and modeling approaches. The study showed that zonated models perform well when geological structures are known. Interpolated heterogeneous fields offer a practical balance between accuracy and data needs. Additional hydrogeological data can reduce uncertainties in complex models. The appropriate level of subsurface detail depends on modeling goals and data availability. These findings suggest that modelers should consider both error types when choosing parameter structures.