Voici les éléments 1 - 6 sur 6
  • Publication
    Accès libre
    The influence of riverbed heterogeneity patterns on river-aquifer exchange fluxes under different connection regimes
    (2017-9)
    Tang, Qi
    ;
    Kurtz, W.
    ;
    ; ;
    Vereecken, H.
    ;
    Hendricks Franssen, Harrie-Jan
    Riverbed hydraulic conductivity (K) is a critical parameter for the prediction of exchange fluxes between a river and an aquifer. In this study, the role of heterogeneity patterns was explored using the fully integrated hydrological model HydroGeoSphere simulating complex, variably saturated subsurface flow. A synthetic 3-D river-aquifer reference model was constructed with a heterogeneous riverbed using nonmulti-Gaussian patterns in the form of meandering channels. Data assimilation was used to test the ability of different riverbed K patterns to reproduce hydraulic heads, riverbed K and river-aquifer exchange fluxes. Both fully saturated as well as variably saturated conditions underneath the riverbed were tested. The data assimilation experiments with the ensemble Kalman filter (EnKF) were carried out for four types of geostatistical models of riverbed K fields: (i) spatially homogeneous, (ii) heterogeneous with multiGaussian distribution, (iii) heterogeneous with non-multi-Gaussian distribution (channelized structures) and (iv) heterogeneous with non-multi-Gaussian distribution (elliptic structures). For all data assimilation experiments, state variables and riverbed K were updated by assimilating hydraulic heads. For saturated conditions, heterogeneous geostatistical models allowed a better characterization of net exchange fluxes than a homogeneous approximation. Among the three heterogeneous models, the performance of non-multi-Gaussian models was superior to the performance of the multi-Gaussian model, but the two tested non-multi-Gaussian models showed only small differences in performance from one another. For the variably saturated conditions both the multi-Gaussian model and the homogeneous model performed clearly worse than the two non-multi-Gaussian models. The two non-multi-Gaussian models did not show much difference in performance. This clearly shows 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.
  • Publication
    Accès libre
    Advances in characterizing surface water-groundwater interactions : combining unconventional data with complex, fully-integrated models
    (2017)
    Schilling, Oliver S.
    ;
    ;
    Caractériser et simuler les interactions entre les eaux de surface et souterraines s’avère être un enjeu de plus en plus important afin de garantir une eau de qualité pour des puits de pompage situés à proximité d’une rivière. Grâce à la dernière génération de modèles numériques physiques intégrant les écoulements de surface et souterrains, il est maintenant possible de simuler tous les processus physiques gouvernant ces interactions. Cependant, les résultats de ces modèles sont souvent peu satisfaisants. Plusieurs études suggèrent que l’utilisation des charges hydrauliques et des débits de la rivière n’est pas suffisante pour décrire et contraindre ces processus complexes. Une revue bibliographique de ces différentes études est présentée dans le chapitre 2. Afin de parvenir à un meilleur calage de ces modèles, l’utilisation d’autres types d’observations que l’on peut qualifier de non-conventionnelles doit être envisagée. Ces observations non-conventionnelles peuvent être par exemple les flux caractérisant les interactions ou la concentration de solutés. Le but principal de cette thèse de doctorat est de montrer que l’utilisation de ces observations non conventionnelles permet non seulement d’améliorer significativement le calage de ces modèles, mais également de réduire grandement les incertitudes de leurs projections.
    Le chapitre 4 présente le développement d’une nouvelle méthode utilisant l’accroissement de cernes des arbres afin de quantifier l’historique de transpiration des arbres riverains. Les observations non-conventionnelles obtenues grâce à cette méthode ont ensuite été utilisées pour caler un modèle HydroGeoSphere couplant les écoulements souterrains et de surface ainsi que la transpiration des végétaux. Une analyse d’incertitude a permis de quantifier la réduction de l’incertitude induite par l’utilisation de ces observations non-conventionnelles.
    Le chapitre 5 présente l’utilisation de traceurs naturels afin de caler un modèle d’écoulements souterrains et de surface. Cette étude fut réalisée dans le cadre d’un essai de traçage réalisé dans un champ de captage situé en Suisse. Après avoir évalué le temps de résidence de l’eau en utilisant les concentrations en argon-37, une analyse du mélange entre les eaux récemment infiltrées et les eaux plus vieilles a été réalisée en analysant les températures d’infiltration des gaz rares. Ces deux informations qui permettent de décrire les interactions entre la rivière et l’aquifère furent utilisées en plus des observations conventionnelles pour caler et contraindre le modèle de ce champ de captage. Les résultats obtenus montrent clairement une amélioration des capacités prédictives du modèle ainsi qu’une diminution de ses incertitudes.
    Le chapitre 3 décrit une nouvelle méthode utilisant la méthode de Monte-Carlos afin d’identifier des zones non saturées entre le lit d’une rivière et d’un aquifère tous deux aux propriétés hétérogènes. Finalement, le chapitre6 présente une nouvelle approche de simulation utilisant HydroGeoSphere, le filtre de Kalman d'ensemble (EnKF) et du nuage informatique (cloud computing) afin d’augmenter les ressources de calcul qui généralement sont importantes dans le cadre de modélisations numériques complexes., The characterization and simulation of the interactions between surface water and groundwater require observations of hydrological state variables and flow processes. While the latest generation of physically-based flow models allows the integrated simulation of all relevant hydrological processes, the current modelling practice is not adequate to provide reliable predictions. Numerous studies suggest that the main reason for this limited predictive capability is that the complex nature of surface water - groundwater systems cannot be sufficiently described and constrained by only considering the ‘classical’ hydrogeological observations of surface water discharge and hydraulic head. An extensive literature review on this topic is provided in Chapter 2. To overcome the problem of inadequate surface water - groundwater flow model calibration, alternative, unconventional observations should be considered, for example observations of solute concentrations or exchange fluxes. With the appropriate modelling and calibration tools, unconventional observations can not only be successfully included in flow model calibration, but by choosing the right tools it is also possible to quantify the information content of unconventional observations towards reducing the predictive uncertainty of flow models. This was the focus of this thesis and is illustrated in multiple studies: In Chapter 4, a new method that uses tree ring growth records to infer the historic transpiration rates of riparian desert trees of the Tarim River was developed. These new and unconventional observations were successfully used for the calibration of an integrated surface water - groundwater - vegetation flow model built with HydroGeoSphere. A post-calibration uncertainty analysis allowed quantifying the high worth of these unconventional observations in reducing the predictive uncertainty of the flow model. In the study presented in Chapter 5, the established tracer methods using Radon-222 and Helium-Tritium were for the first time complemented by a novel tracer method based on Argon-37, which allowed closing a previously existing gap in residence times characterization. A multi-tracer study carried out on an important drinking water wellfield in Switzerland provided an ideal framework to test the new method alongside other tracer methods. Following the successful characterization of residence times of groundwater in the drinking water wellfield, a quantification of mixing of different types of groundwater, i.e. of recently infiltrated river water and of old groundwater, was achieved through noble gas end-member analysis. The information on mixing of different types of groundwater was subsequently used to inform the parametrization of an integrated surface water - groundwater flow model of the drinking water wellfield. It could be shown that the use of mixing information contains information about exchange fluxes and can successfully inform a flow model parametrization beyond the capabilities of classical observations of groundwater heads and surface water discharge.
    One important aspect of surface water - groundwater interactions is the state of connection between the two water bodies. If unsaturated flow processes are expected in a given surface water - groundwater system, it is important that the numerical model which is used to simulate these systems is capable of simulating unsaturated flow. The study in Chapter 3 is dedicated to the development of a Monte-Carlo-based method which allows the rapid quantification of the potential for unsaturated flow processes underneath heterogeneous riverbeds overlying heterogeneous aquifers. This in turn allows a preliminary assessment of the conceptual model of a give surface water-groundwater system.
    And finally, as the integrated simulation of surface water - groundwater interactions requires a lot of computational resources, especially when unsaturated flow processes dominate, a modelling framework using HydroGeoSphere, Ensemble Kalman Filter and cloud resources was developed in order to leverage typically limited computational resources. This framework is presented in Chapter 6.
  • Publication
    Accès libre
    Tutorials as aflexible alternative to GUIs: An example for advanced model calibration using Pilot Points
    Environmental modelling software with graphical user interfaces (GUIs) are user friendly and help tofocus on the aspects of modelling rather than on the technicality of the underlying code. This is afundamental advantage compared to codes without a GUI, as the absence of a GUI can prohibit thewidespread application of a software tool. However, there is a downside to GUIs, too. They commonly lagbehind the newest software development. We argue that tutorials are aflexible but undervaluedalternative to GUIs, and are convinced that tutorials can help to make complex software accessible to anincreased number of users. As an example to demonstrate our point, we have written a tutorial thatillustrates the use of Pilot-Point based calibration made available through PEST in the HydroGeoSpheremodelling environment. We hope that this example will encourage the modelling community to developtutorials and make them available to the wider public.
  • Publication
    Accès libre
    Tutorials as a flexible alternative to GUIs: An example for advanced model calibration using Pilot Points
    Environmental modelling software with graphical user interfaces (GUIs) are user friendly and help to focus on the aspects of modelling rather than on the technicality of the underlying code. This is a fundamental advantage compared to codes without a GUI, as the absence of a GUI can prohibit the widespread application of a software tool. However, there is a downside to GUIs, too. They commonly lag behind the newest software development. We argue that tutorials are a flexible but undervalued alternative to GUIs, and are convinced that tutorials can help to make complex software accessible to an increased number of users. As an example to demonstrate our point, we have written a tutorial that illustrates the use of Pilot-Point based calibration made available through PEST in the HydroGeoSphere modelling environment. We hope that this example will encourage the modelling community to develop tutorials and make them available to the wider public.
  • Publication
    Accès libre
    A hydraulic mixing-cell method to quantify the groundwater component of streamflow within spatially distributed fully integrated surface water–groundwater flow models
    (2011) ; ;
    Craig T. Simmons
    ;
    Therrien, René
    ;
    Werner, A.D.
    ;
    Dandy, G.C.
    ;
    Maier, H.R.
    The complexity of available hydrological models continues to increase, with fully integrated surface water–groundwater flow and transport models now available. Nevertheless, an accurate quantification of streamflow generation mechanisms within these models is not yet possible. For example, such models do not report the groundwater component of streamflow at a particular point along the stream. Instead, the groundwater component of streamflow is approximated either from tracer transport simulations or by the sum of exchange fluxes between the surface and the subsurface along the river. In this study, a hydraulic mixing-cell (HMC) method is developed and tested that allows to accurately determine the groundwater component of streamflow by using only the flow solution from fully integrated surface water–groundwater flow models. By using the HMC method, the groundwater component of streamflow can be extracted accurately at any point along a stream provided the subsurface/surface exchanges along the stream are calculated by the model. A key advantage of the HMC method is that only hydraulic information is used, thus the simulation of tracer transport is not required. Two numerical experiments are presented, the first to test the HMC method and the second to demonstrate that it quantifies the groundwater component of streamflow accurately.
  • Publication
    Accès libre
    Evaluation of outputs from automated baseflow separation methods against simulated baseflow from a physically based, surface water-groundwater flow model
    Partington, D
    ;
    ;
    Simmons, C. T
    ;
    Werner, A. D
    ;
    Therrien, R
    ;
    Maier, H. R
    ;
    Dandy, G. C
    Summary Baseflow is often considered to be the groundwater discharge component of streamflow. It is commonly estimated using conceptual models, recursive filters or a combination of the two. However, it is difficult to validate these methods due to the current challenges of measuring baseflow in the field. In this study, simulation of a synthetic catchment’s response to rainfall is carried out using a fully integrated surface water-groundwater flow model. A series of rainfall events with differing recovery periods and varied antecedent moisture conditions is considered to span a range of different streamflow generation dynamics. Baseflow is estimated for the outlet hydrograph of the synthetic catchment using a selection of commonly used automated baseflow separation methods. These estimates are compared to the baseflow signal obtained from the numerical model, which serves as the control experiment. Results from these comparisons show that depending on the method used, automated baseflow separation underestimates the simulated baseflow by as much as 28%, or overestimates it by up to 74%, during rainfall events. No separation method is found to be clearly superior to the others, as the performance of the various methods varies with different soil types, antecedent moisture conditions and rainfall events. The differences between the various approaches clearly demonstrate that the baseflow separation methods investigated are not universally applicable.