River-aquifer interaction and the role of streambeds

A proper characterization of river-aquifer exchange fluxes is important for several reasons. It is relevant for groundwater management in order to determine the regional water balance (infiltrating/exfiltrating rivers strongly influence the amount of available water in an aquifer) and because infiltrating, contaminated stream water (e.g., bacteria, antibiotics, pesticides, industrial accidents) is a potential danger for groundwater quality. An accurate assessment of river-aquifer interactions can also be important for ecological corridors because exchange fluxes have a relatively large impact on the amount of river discharge under drought conditions. In order to assess whether a minimum river discharge, needed for the ecology close to the river, is maintained under drought conditions, it is essential to know the interactions between streams and groundwater. The prediction of the exchange fluxes might also be important for flood prediction, as especially during floods large amounts of water from the stream can infiltrate into aquifers beneath the stream. This can reduce the river discharge significantly, especially for long river reaches.
The properties of streambeds together with the hydraulic gradients determine the fluxes of water, energy and solutes between streams and aquifers. Little permeable streambeds (e.g., streambeds which are clogged with a layer of very fine material) will result in smaller exchange fluxes. Highly permeable streambeds (e.g., stony streambeds in steeper streams with regular flooding events) exchange potentially more water, solutes and energy with the adjacent aquifer. A problem is that the material properties of streambeds are strongly heterogeneous in space and also in time. Flood events change streambeds and cause erosions in some river sections (resulting in more permeable riverbeds) and deposit large amounts of fine materials elsewhere (resulting in less permeable riverbeds). However, it is very difficult to characterize these erosion/sedimentation events deterministically.
The main findings from our research are:

1. Heterogeneity of riverbeds can play an important role for river-aquifer exchange fluxes. Replacing a heterogeneous riverbed with effective, homogeneous parameters can often not reproduce adequately river-aquifer exchange fluxes. Irvine et al. (2012) found that effective spatially homogeneous parameters, fitted for a certain flow regime (i.e., fully connected, disconnected or partially connected conditions), result in erroneous flux estimates if the flow regime is changed. However, if the flow regime does not change (for example, river-aquifer are fully connected in both cases) the homogeneous approximation introduces only a small error. Given the importance of the connection regime and changes of the connection regime, Schilling et al. (2017) developed a criterion to predict the extent of unsaturated zones which develop beneath riverbeds. This criterion includes heterogeneity of both riverbeds and aquifer.
2. Given the importance of spatial heterogeneity of riverbeds, it was tested to which degree this heterogeneity can be characterized by piezometric head data. It was found that saturated hydraulic conductivity of heterogeneous streambeds is best inversely estimated by a high-resolution representation of the streambed (stochastic field approach). It results in the best characterization of states and fluxes in a model for river-groundwater interaction. If the streambed properties are very coarsely represented with help of effective parameters, the net exchange fluxes over longer time periods along the stream are systematically underestimated (although the measured heads are reproduced well) (Kurtz et al., 2013).
3. Further research focused on the role of groundwater temperature data to characterize heterogeneous riverbeds. These data were used in a data assimilation scheme linked to a coupled groundwater flow-heat transport model including river-aquifer interaction. The upper Limmat valley in Zurich (Switzerland) was the case study. It was found that groundwater temperature data could only contribute marginally to further improve the characterization of the river-aquifer interaction, but the groundwater temperature were important to better model heat transport (Kurtz et al., 2014).
4. More recent work started to look at the role of more complicated spatial patterns of riverbed properties on river-aquifer exchange fluxes. First, the interaction was studied with the simulation model SPRING, and later with the fully integrated hydrological model HydroGeoSphere simulating complex, variably saturated subsurface flow. In this work it was found again that characterizing heterogeneity of riverbed K is important. Moreover, particularly under variably saturated flow conditions the mean and the variance of riverbed K do not provide enough information for exchange flux characterization and additional histogram information of riverbed K provides crucial information for the reproduction of exchange fluxes (Tang et al., 2015, 2017).

Besides spatial variability of riverbeds, also temporal variability of riverbeds has been subject of study. For the river Limmat close to the city of Zurich (Switzerland) it was found that sudden flood events significantly change streambed permeabilities. It was also found that seasonal temperature variations have a significant influence on streambed permeabilities (Doppler et al., 2017). Taking these variations into account with a coupled groundwater flow-heat transport modeling approach with temperature dependent hydraulic conductivities improves modeling results (Engeler et al., 2011). A synthetic study revealed that with piezometric head measurements close to the stream it is possible to detect changes in riverbed permeability over time. This can be done with data assimilation methods that update model states and parameters in real-time. A disadvantage is that the model shows a delayed response with respect to sudden changes in the riverbed properties. This can be partly corrected by the use of inflation methods in combination with Ensemble Kalman Filtering (Kurtz et al., 2012). Tang et al. (2018) demonstrated that also for a real-world case, the severe flooding event in the Emme river in Switzerland, updating riverbed topography and riverbed properties in a data assimilation procedure improved the modelling of river discharge and groundwater flow near the river.

Contact person:

Harrie-Jan Hendricks Franssen
Agrosphere (IBG-3)
Forschungszentrum Jülich GmbH
Leo Brandtstrasse
52425 Jülich
Germany
Tel. 02461 / 61-4462
E-mail: h.hendricks-franssen@fz-juelich.de

Link to project partners:

University of Neuchatel, Switzerland

Link to former project partners:

Institute for Environmental Engineering, ETH Zürich, Switzerland

TK Consult AG, Zürich, Switzerland

Relevant publications:

Schilling, O., D.J. Irvine, H.J. Hendricks Franssen, and P. Brunner, 2017. Estimating the spatial extent of unsaturated zones in heterogeneous river-aquifer systems. Water Resources Research, accepted.

Tang, Q., W. Kurtz, O.S. Schilling, P. Brunner, H. Vereecken, and H.J. Hendricks Franssen. 2017. The influence of riverbed heterogeneity patterns on river-aquifer exchange fluxes under different connection regimes. An evaluation with an integrated hydrological model and data assimilation. Journal of Hydrology, 554, 383-396.

Brunner, P., R. Therrien, P. Renard, C.T. Simmons, and H.J. Hendricks Franssen, 2017. Advances in understanding river-groundwater interactions. Reviews of Geophysics, 55, doi: 10.1002/2017RG000556.

Tang, Q., W. Kurtz, P. Brunner, H. Vereecken, and H.J. Hendricks Franssen. 2015. Characterization of river-aquifer exchange fluxes: the role of spatial patterns of riverbed hydraulic conductivities. Journal of Hydrology 531, 111-123.

Kurtz, W., H.J. Hendricks Franssen, P. Kaiser, and H. Vereecken. 2014. Joint assimilation of piezometric heads and groundwater temperatures for improved modeling of river-aquifer interactions. Water Resources Research 50, doi:10.1002/2013WR014823.

Kurtz, W., H.J. Hendricks Franssen, P. Brunner, and H. Vereecken. 2013. Is high resolution inverse characterization of heterogeneous river bed hydraulic conductivity needed and possible? HESS 17, 3795-3813.

Kurtz, W., H.J. Hendricks Franssen, and H. Vereecken. 2012. Identification of time-variant river bed properties with the Ensemble Kalman Filter. Water Resources Research, 48, W10534, doi:10.1029/2011WR011743.

Irvine, D., P. Brunner, H.J. Hendricks Franssen and C.T. Simmons. 2012. Heterogeneous or homogeneous? Implications of simplifying heterogeneous streambeds in models of losing streams. Journal of Hydrology 424-425, 16-23. doi: 10.1016/j.jhydrol.2011.11.051.

Hendricks Franssen, H.J., H.P. Kaiser, U. Kuhlmann, G. Bauser, F. Stauffer, R. Müller and W. Kinzelbach. 2011. Operational real-time modeling with EnKF of variably saturated subsurface flow including stream-aquifer interaction and parameter updating. Water Resources Research 47, W02532, doi:10.1029/2010WR009480.

Engeler, I., H.J. Hendricks Franssen, R. Müller and F. Stauffer. 2011. The importance of coupled modelling of variable saturated groundwater flow-heat transport for assessing river-aquifer interactions. Journal of Hydrology 397, 295-305, doi:10.1016/j.hydrol.2010.12.007.

Doppler, T., H.J. Hendricks Franssen, H.P. Kaiser, U. Kuhlmann and F. Stauffer. 2007. Field evidence of a dynamic leakage coefficient for modelling river-aquifer interactions. Journal of Hydrology 347, 177-187..doi: 10.1016/j.jhydrol.2007.09.017.