VTraFlux − A model toolbox to determine transient vertical exchange fluxes in hyporheic sediments from time series of natural tracers

• Published in: Journal of hydrology (Amsterdam) Vol. 651; p. 132520
• Main Authors: Schaper, Jonas L., Zarfl, Christiane, Lewandowski, Joerg, Cirpka, Olaf A.
• Format: Journal Article
• Language: English
• Published: Elsevier B.V 01.04.2025
• Subjects: algorithms; data collection; electrical conductivity; equations; groundwater; Groundwater – surface water exchange; heat; Heat as a tracer; hyporheic zone; Parameter inference; porosity; spectral analysis; surface water; temperature; thermal conductivity; time series analysis; Transient flow; Transport simulation; uncertainty; Groundwater – surface water exchange; Transient flow; Transport simulation; Heat as a tracer; Parameter inference
• DOI: 10.1016/j.jhydrol.2024.132520
• ISSN: 0022-1694

•Model toolbox to estimate time-varying river to riverbed travel times.•Toolbox uses multiple natural tracers such as heat and electrical conductivity (EC)•L-curve method used to optimize Tikhonov regularization.•Joint inversion of EC and temperature reduces uncertainty in parameter estimates. Natural tracers such as heat and electrical conductivity (EC) are widely used to estimate the direction and magnitude of exchange flows between surface water and groundwater. As a consequence, a variety of model approaches and toolboxes exists that derive fluxes from time series of fluctuations of EC and temperature. However, the majority of models performs relatively poorly when hydrological conditions are subject to sudden or regular changes and are unable to resolve resulting flux transients, particularly in situations where fluxes change on a sub-daily time scale. Moreover, few model toolboxes jointly use EC and heat as tracers to estimate water exchange flows across surface water – groundwater interfaces. In the present study, we introduce VTraFlux, a collection of model routines, allowing users to estimate transient exchange flows between surface water and groundwater from multiple natural-tracer time series at a user-defined temporal resolution. The model is based on the advection–dispersion equation for EC and with retardation for temperature, uses the L-curve method to optimize Tikhonov regularization, and the differential evolution adaptive metropolis algorithm (DREAM) to infer unknown parameters. We demonstrate its applicability with three datasets that were collected under transient hydrological conditions differing with respect to the dominant exchange flow direction. The estimated exchange flows were compared with results obtained by VFLUX 2.0, a model toolbox that relies on spectral analysis and an analytical solution of the heat-transport equation. The flux transients obtained by VTraFlux are more pronounced compared to VFLUX-derived estimates, particularly under conditions where rivers gained water by groundwater inflow. Jointly inverting EC and temperature time series reduces the uncertainty of the estimated thermal dispersivity, porosity, volumetric heat capacity, thermal conductivity, and apparent porewater velocity time series. The open-source model is scripted in Python, thus enabling users to adapt and modify the code.