Assessing the water balance of the Upper Rhine Graben hydrosystem

SummaryThe Upper Rhine alluvial aquifer is an important transboundary water resource. However, as in many alluvial systems, the aquifer inflows and outflows are not precisely known because of the difficulty of estimating the river infiltration flux and the boundary subsurface flow. To provide a thorough representation of the aquifer system, a coupled surface–subsurface model was applied to the whole aquifer basin, and several parameter sets were tested to investigate the uncertainty due to poorly known parameters (e.g. aquifer transmissivity computed by an inverse model, river bed characteristics). Twelve simulations were run and analyzed using standard statistical criteria and also a more advanced statistical method, the Karhunen Loève transform (KLT). This analysis showed that, although the model performed reasonably well, some piezometric level underestimations persisted in the south of the basin. An accurate representation of the aquifer behaviour would require river infiltration and the functioning of irrigation canals in the Hardt area to be taken into account. It also appeared that increasing the maximum river infiltration flow deteriorated the quality of the results.River infiltration to the aquifer was estimated to represent about 80% of the aquifer inflows with a mean annual value around 115 ± 16.5 m3/s, thus with an uncertainty of 14%. This quantity is larger than estimated in previous studies but is in agreement with some results obtained during low water periods. This important conclusion highlights the vulnerability of the Upper Rhine Graben aquifer to pollution from the rivers and to climate change since it is highly probable that the rivers’ regimes will be affected by reduced snow cover on the neighbouring mountain ranges.

► We build a regional coupled model for the Upper Rhine hydrosystem.
► Parameter sensitivity analysis is undertaken with classic and more advanced methods.
► Aquifer recharge and discharge water budgets are assessed quantitatively.