[A-HW19] 水循環・水環境

[AHW19-02] Evaluating the worth of environmental tracers in alluvial stream-aquifer systems

*Hugo Delottier¹、John Doherty²、Morgan Peel¹、Philip Brunner¹ (1.Centre d'Hydrogéologie et de Géothermie, University of Neuchâtel, 2000 Neuchâtel, Switzerland、2.Watermark Numerical Computing, Corinda, QLD)

キーワード：Data worth, Environmental tracers, Integrated Surface and Subsurface Hydrological Model

Alluvial stream-aquifer systems are a preferred source of water supply due to the accumulation of millennia of channel meandering, aggregation and erosional flow processes that favour heterogeneous systems with preferential flow paths. In these environments, groundwater levels are usually close to the surface, making these aquifers a convenient and easily exploitable, but also vulnerable, groundwater resource. Sustainable management of surface water and groundwater resources in alluvial environments therefore requires a sound understanding of aquifer heterogeneity and exchange fluxes between surface water and groundwater. Physics-based numerical models can help to inform about these complex processes and their predictive power is a major advantage to support model management with key predictions such as groundwater residence times and mixing ratios. However, classical observations such as hydraulic heads may not be sufficient to constrain all the model predictions and the assimilation of complementary observations such as environmental tracer concentrations may have great potential when included in the history matching process. However, their potential and how to combine them to maximise uncertainty reduction remains unclear. We propose to evaluate the value of environmental tracer concentrations based on a synthetic yet realistic stream-aquifer system, where groundwater concentrations of ²²²Rn, ⁴He and ³⁷Ar can be considered as complementary observations to represent all relevant timescales in alluvial systems. ALLUVSIM, a pseudo-process-based software capable of generating typical alluvial architectures, will be used to provide realistic categorical structures with specific hydraulic properties such as hydraulic conductivity and porosity. HydroGeoShere (HGS), an Integrated Surface and Subsurface Hydrologic Model (ISSHM), will be used to explicitly simulate tracer concentrations. The synthetic experiment will be conducted over a daily transient reduced pumping test, where significant stress will be induced in the system to maximise the potential of the calibration dataset to inform predictions. Data worth will be systematically quantified by applying Data Space Inversion with different realistic alluvial architectures, potentially covering a wide range of complexity levels.