Abstract
Riverbed temperature profiles are frequently used to estimate vertical river–aquifer exchange fluxes. Often in this approach, strictly vertical flow is assumed. However, riverbeds are heterogeneous structures often characterised by complex flow fields, possibly violating this assumption. We characterise the meter-scale variability of river–aquifer interaction at two sections of the Aa River, Belgium, and compare vertical flux estimates obtained with a 1D analytical solution to the heat transport equation with fluxes simulated with a 3D groundwater model (MODFLOW) using spatially distributed fields of riverbed hydraulic conductivity. Based on 115 point-in-time riverbed temperature profiles, vertical flux estimates that are obtained with the 1D solution are found to be higher near the banks than in the center of the river. The total exchange flux estimated with the 3D groundwater model is around twice as high as the estimate based on the 1D solution, while vertical flux estimates from both methods are within a 10% margin. This is due to an important contribution of non-vertical flows, especially through the riverbanks. Quasi-vertical flow is only found near the center of the river. This quantitative underestimation should be considered when interpreting exchange fluxes based on 1D solutions. More research is necessary to assess conditions for which using a 1D analytical approach is justified to more accurately characterise river–aquifer exchange fluxes.
Highlights
All recorded temperature profiles show a decrease in temperature with depth, which is consistent with higher temperatures of surface water compared to groundwater during summer (Figure 3)
Analytical solution are negative, indicating upward flow towards the river, i.e., the river is gaining over the observed period
Vertical flux estimates range from −28.9 mm/day to
Summary
The novelty of our study lies in the combination of highresolution riverbed hydraulic conductivity measurements with 3D groundwater modelling to simulate complex flow patterns in the riverbed at the meter-scale. These high-resolution flows are compared with exchange fluxes estimated with a 1D heat tracer method. We compare these spatially distributed flux estimates to estimates obtained from a numerical 3D groundwater model (set up in MODFLOW) that essentially quantifies fluxes based on Darcy’s law This model makes use of previously obtained high-resolution data on riverbed hydraulic conductivity [51] to simulate riverbed heterogeneity
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have