Do observation wells in fractured porous aquifers measure water head in the fracture network, water head in the matrix, or some combination of both? This question necessarily arises when calibrating dual‐continuum hydrogeological models against on‐field data. One can assume that observation wells measure fracture water head, because matrix permeability is negligible compared to fracture permeability. Nevertheless, this reasoning is invalid for wells poorly‐connected to the fractures. Yet, the possibility of such a poor connection at given depths has never been implemented in a physics‐based manner when comparing matrix and fracture water heads simulated by dual‐continuum models to on‐field data. To fill this knowledge gap, a physically based, easy to calibrate, open‐source postprocessing tool, POWeR‐FADS (Program for Observation Well Representation in Fractured Aquifer Dual‐continuum Simulations), available at https://github.com/BJeannot1/POWeR-FADS , has been developed. It introduces as parameters well geometry and the altitude of lowest interception of the fractures by the well. From these, POWeR‐FADS nonintrusively postprocesses time series of matrix and fracture water heads at the well, as simulated by any planar, bidimensional dual‐continuum hydrogeological model, to calculate water exchanges involving the observation well and thus the evolution of water level in the well. Synthetic test cases show that POWeR‐FADS makes it possible to simulate peculiar behaviors that are similar to patterns actually observed by the authors in on‐site observation wells of a fractured porous aquifer, like “floors” in observed water levels, delayed but sharp rises at the beginning of recharge events, or inflexion points accelerating the drawdown velocity during the recession phase.

Abstract. Understanding the variability of the chemical composition of surface waters is a major issue for the scientific community. To date, the study of concentration–discharge relations has been intensively used to assess the spatiotemporal variability of the water chemistry at watershed scales. However, the lack of independent estimations of the water transit times within catchments limits the ability to model and predict the water chemistry with only geochemical approaches. In this study, a dimensionally reduced hydrological model coupling surface flow with subsurface flow (i.e., the Normally Integrated Hydrological Model, NIHM) has been used to constrain the distribution of the flow lines in a headwater catchment (Strengbach watershed, France). Then, hydrogeochemical simulations with the code KIRMAT (i.e., KInectic Reaction and MAss Transport) are performed to calculate the evolution of the water chemistry along the flow lines. Concentrations of dissolved silica (H4SiO4) and in basic cations (Na+, K+, Mg2+, and Ca2+) in the spring and piezometer waters are correctly reproduced with a simple integration along the flow lines. The seasonal variability of hydraulic conductivities along the slopes is a key process to understand the dynamics of flow lines and the changes of water transit times in the watershed. The covariation between flow velocities and active lengths of flow lines under changing hydrological conditions reduces the variability of water transit times and explains why transit times span much narrower variation ranges than the water discharges in the Strengbach catchment. These findings demonstrate that the general chemostatic behavior of the water chemistry is a direct consequence of the strong hydrological control of the water transit times within the catchment. Our results also show that a better knowledge of the relations between concentration and mean transit time (C–MTT relations) is an interesting new step to understand the diversity of C–Q shapes for chemical elements. The good match between the measured and modeled concentrations while respecting the water–rock interaction times provided by the hydrological simulations also shows that it is possible to capture the chemical composition of waters using simply determined reactive surfaces and experimental kinetic constants. The results of our simulations also strengthen the idea that the low surfaces calculated from the geometrical shapes of primary minerals are a good estimate of the reactive surfaces within the environment.

Magnetic Resonance Sounding (MRS) measurements are acquired at 16 stations in the Strengbach headwater catchment (Vosges Mountains – France). These data, rendering the vertical distribution of water contents in the subsurface, are used to show their potential in conditioning a hydrological model of the catchment, as described in the article “Magnetic resonance sounding measurements as posterior information to condition hydrological model parameters: Application to a hard-rock headwater catchment” – Journal of Hydrology (2020). Acquisition protocols follow a free induction decay scheme. Data are filtered by applying a band-pass filter at the Larmor frequency. A filter removing the 50 Hz noise is also applied with the exception of data at a Larmor frequency close to the 50 Hz harmonic. The signal envelopes are then fitted by a decaying exponential function over time to estimate the median characteristic relaxation time of each MRS sounding.

In headwater catchment, the calibration of hydrological models is complex due to the scarcity of data in mountainous areas. Here, an innovative methodology is developed to condition hydrological model parameters by using magnetic resonance sounding (MRS) measurements in combination with stream flow rate data. MRS has the specificity in the various geophysical imaging techniques of being mainly sensitive to the vertical distribution of water content among the subsurface. In a way very similar to hydraulic head observations, these local distributions of water content may serve as information in a hydrological model to pattern subsurface flow by seeking model parameters. Simulations are run with different sets of parameters of a hydrological model. Each simulation provides as an output a 4-D map (3-D spatial plus time) of the vertical water content distributions over the whole catchment and their fluctuations over time. This output is then used to simulate the MRS signal that would be produced by the estimated water content. The simulated MRS signal is compared to measured MRS data to determine which hydrological simulations (which model parameters) are close to observations. The approach is applied on a hard-rock headwater catchment housing a very shallow and thin aquifer where an MRS survey covers the whole studied site. Hydraulic parameters of an integrated hydrological model of the catchment are spatially distributed by zones with uniform values, the prior delineation of the zones being guided by pedological studies. As MRS measurements supply local but spatially distributed information, the method conditions the various zones on their parameter values in a much better way than the classical (in headwater catchments) measure of the stream flow rate at the outlet of the system. Finally, hydrological simulation and time-dependent MRS forward calculations can help identifying possible locations for MRS stations to monitor the transient behavior of the hydrological state of the catchment.

In mountainous area, spring water constitutes the only drinking water resource and local economy is highly dependent on forest health and productivity. However, climate change is expected to make extreme water shortage episodes more and more frequent. Forest is therefore more and more exposed to water stress. It appears necessary to quantify the drought induced by water deficit to evaluate forest vulnerability and to plan the future of forest management. In this study we quantified the 2018 water deficit experienced by the forest in the Strengbach catchment, located in the French Vosges mountains. Three methods for estimating catchment water storage changes (WSC) have been compared. The first relies on superconducting gravimeter monitoring while the second relies on catchment water balance. The third one relies on global hydrological model MERRA2. We show that WSC estimated from measured gravity changes correlate well with WSC estimated from catchment water balance while WSC inferred from MERRA2 significantly differs. The Strengbach catchment water cycle is mostly annual but exhibits significant interannual variability associated with the 2018 drought episode: August 2018 has a water deficit of 37 mm (as inferred from catchment water balance) or 76 mm (as seen with superconducting gravimetry) compared to August 2017. We illustrate here the use of superconducting gravimeter monitoring as an independent proxy for WSC in a mountainous catchment while most of hydro-gravimetric studies have been conducted on relatively flat areas. We therefore contribute to expand the area of use of high precision gravity monitoring for the hydrological characterization of the critical zone in mountainous context. This innovative method may help to assess forest vulnerability to drought in the context of climate change.

Rohrschollen Island is an artificial island of the large Upper Rhine river whose geometry and hydrological dynamics are the result of engineering works during the 19th and 20th centuries. Before its channelization, the Rhine river was characterized by an intense hydromorphological activity which maintained a high level of biodiversity along the fluvial corridor. This functionality considerably decreased during the two last centuries. In 2012, a restoration project was launched to reactivate typical alluvial processes, including bedload transport, lateral channel dynamics, and surface–subsurface water exchanges. An integrated hydrological model has been applied to the area of Rohrschollen Island to assess the efficiency of the restoration regarding surface and subsurface flows. This model is calibrated using measured piezometric heads. Simulated patterns of water exchanges between the surface and subsurface compartments of the island are checked against the information derived from thermal infrared (TIR) imaging. The simulated results are then used to better understand the evolutions of the infiltration–exfiltration zones over time and space and to determine the physical controls of surface–subsurface interactions on the hydrographic network of Rohrschollen Island. The use of integrated hydrological modeling has proven to be an efficient approach to assess the efficiency of restoration actions regarding surface and subsurface flows.

The temporal variability of transit-time distributions (TTDs) and residence-time distributions (RTDs) in hydrological systems has received particular attention recently because of their ability to inform on elementary processes impacting geochemical signatures and water fluxes in ecosystems. To date, these distributions and their temporal variability have been mainly investigated through concentration measurements of conservative geochemical or isotopic tracers. Even though physically-based and distributed hydrological models can render interpretations of TTDs/RTDs in terms of processes and physical controls, the variability of TTDs and RTDs has barely been studied using distributed hydrological modeling. In this study, an integrated hydrological model has been coupled with particle tracking algorithms and applied to the Strengbach Catchment – a small mountainous catchment belonging to the French network of critical zone observatories – to investigate the eventual link between water storage in the catchment and the temporal variability of TTDs and RTDs. The model calibration is performed relying upon both classical streamflow measurements and magnetic resonance sounding, a geophysical measure sensible to the water content in the subsurface. The model is then run over a 10-year period for which time distributions are calculated at various deadlines. The results show that the response of the Strengbach catchment is uncommon with short mean transit times (approximately 150-200 days) and a weak variability of TTDs and RTDs with the water storage. This specific behavior is mainly linked to the small size of the system and specific climatic and topographic conditions. Because the hydrological model was calibrated on the basis of unusual data (local water contents inferred via MRS measurements), ongoing investigations target the evaluation of the sensitivity of transit time distributions with respect to uncertainties plaguing calibrating data.

L’objectif du travail de thèse est d'œuvrer au développement et à l’application d’un modèle hydrologique intégré pré-existant (Pan et al., 2015; Weill et al., 2017) : Normally Integrated Model (NIM). La spécificité de ce modèle est d’intégrer l’équation d’écoulement souterrain 3D sur la direction perpendiculaire au substratum, de façon à se ramener à un problème en deux dimensions. Il en résulte un gain substantiel en termes de temps de simulation, et une économie du point de vue de l’espace mémoire requis. Dans le cadre de cette thèse, NIM a été entièrement recodé et optimisé. Un module de ruissellement 2-D a également été implémenté, ce qui a permis d’appliquer le modèle en situations réelles sur deux bassins versants distincts. En sus, la validité des simulations issues de NIM et l'efficacité du modèle en termes de temps de calcul ont été évaluées sur de nombreux cas tests synthétiques.

An integrated hydrological model that couples a low-dimensional (2-D) subsurface model to 1-D river flow and 2-D overland flow is presented and applied to a real-world system. The 2-D subsurface equation stems from a careful integration a 3-D Richards equation along a local direction normal to the bottom of the aquifer. A Normally Integrated Model (NIM) is first compared with the integrated model CATHY (CATchment HYdrology) in regard to a synthetic test case with complex geometry, heterogeneous hydraulic conductivity field of the subsurface compartment, and various interactions between the surface and the subsurface. This exercise reveals the gain in computation time associated with the use of a low-dimensional model (by a factor of 6). NIM is then applied to simulate the hydrodynamics of an actual restored hydrosystem, namely, Rohrschollen Island of the Upper Rhine River, with the aim of assessing the ability to address highly transient hydrological conditions and the efficiency of the restoration strategy. The calibration and validation steps associated with this exercise show that the coupled model describes the hydraulic head evolution and water mass balance induced by artificial injections of surface water in a channel dug for the purpose of reconnecting the Rhine to an old and clogged channel. The low-dimensional coupled model of this study is simpler than the state-of-the-art integrated models employing the solution of the 3-D Richards equation. It is, however, suited to investigate fine-scale problems in relation to water resource management and hydraulic developments, with reasonable computational costs and a physically-based approach to model flow processes.