numerical modelling; hydrogeophysics; ensemble modelling; geostatistics; alluvial deposits; inverse methods; error analysis; model simplification; kriging; optimization; Braided rivers; Heterogeneity; Hydrogeology; Geostatistics; Geophysics; Uncertainty Quantification; Model reduction; Sedimentology
Zahner Tobias, Lochbühler Tobias, Mariethoz Grégoire, Linde Niklas (2016), Image synthesis with graph cuts: a fast model proposal mechanism in probabilistic inversion, in
Geophysical Journal International, 204(2), 1179-1190.
Josset Laureline, Ginsbourger David, Lunati Ivan (2015), Functional error modeling for uncertainty quantification in hydrogeology, in
Water Resources Research , 51(2), 1050-1068.
Linde Niklas, Renard Philippe, Mukerji Tapan, Caers Jef (2015), Geological realism in hydrogeological and geophysical inverse modeling: A review, in
Advances in Water Resources, 86, 86-101.
Lochbühler Tobias, Vrugt Jasper A., Sadegh Mojtaba, Linde Niklas (2015), Summary statistics from training images as prior information in probabilistic inversion, in
Geophysical Journal International, 201, 157-171.
Linde Niklas, Lochbühler Tobias, Dogan Mine, Van Dam Remke L., Van Dam Remke L. (2015), Tomogram-based comparison of geostatistical models: Application to the Macrodispersion Experiment (MADE) site, in
Journal of Hydrology, 531, 543-556.
Lochbühler Tobias, Pirot Guillaume, Straubhaar Julien, Linde Niklas (2014), Conditioning of Multiple-Point Statistics Facies Simulations to Tomographic Images, in
Mathematical Geosciences, 46, 625-645.
Linde Niklas (2014), Falsification and corroboration of conceptual hydrological models using geophysical data, in
WIREs Water, 1(2), 1-21.
Scheidt Céline, Renard Philippe, Caers Jef (2014), Prediction-focused subsurface modeling: Investigating the need for accuracy in flow-based inverse modeling, in
Mathematical Geosciences, 47(2), 173-191.
Lochbuehler Tobias, Breen Stephen J., Detwiler Russell L., Vrugt Jasper A., Linde Niklas (2014), Probabilistic electrical resistivity tomography of a CO2 sequestration analog, in
Journal of Applied Geophysics, 107, 80-92.
Fenwick Darryl, Scheidt Céline, Caers Jef (2014), Quantifying Asymmetric Parameter Interactions in Sensitivity Analysis: Application to Reservoir Modeling, in
Mathematical Geosciences, 46(4), 493-511.
Pirot Guillaume, Straubhaar Julien, Renard Philippe (2014), Simulation of braided river elevation model time series with multiple-point statistics, in
Geomorphology, 214, 148-156.
Meerschman Eef, Pirot Guillaume, Mariethoz Gregoire, Straubhaar Julien, Meirvenne Marc Van, Renard Philippe (2013), A practical guide to performing multiple-point statistical simulations with the Direct Sampling algorithm, in
Computers & Geosciences, 52, 307-324.
Ginsbourger David, Rosspopoff Bastien, Pirot Guillaume, Durrande Nicolas, Renard Philippe (2013), Distance-based kriging relying on proxy simulations for inverse conditioning, in
Advances in Water Resources, 52, 275-291.
Josset Laureline, Lunati Ivan (2013), Local and Global Error Models to Improve Uncertainty Quantification, in
Mathematical Geosciences, 45(5), 601-620.
Lochbühler Tobias, Doetsch Joseph, Brauchler Ralf, Linde Niklas (2013), Structure-coupled joint inversion of geophysical and hydrological data, in
Geophysics, 78(3), ID1-ID14.
Schmelzbach C., Huber E., Efficient deconvolution of ground-penetrating radar data., in
IEEE Transactions on Geoscience and Remote Sensing.
Scheidt Céline, Tahmasebi Pejman, Pontiggia M, Da Pra A, Caers Jef, Updating joint uncertainty in trend and depositional scenario for reservoir exploration and early appraisal, in
Computational Geosciences.
Numerical models are the only tools that allow forecasting the behavior of complex hydrological systems. However, the current state of the art suggests major shortcomings in our ability to understand and predict such systems. Recent developments in quantitative hydrology as well as numerical and stochastic modeling open new possibilities for addressing some of the most important shortcomings. However, these recent advances have not yet been integrated in a unique modeling framework. In this project, we use this untapped potential by building a common numerical framework based on the latest advances in quantitative geology, mathematics, and physics. This will enable stochastic modeling with unprecedented levels of realism and purpose. Our objectives are motivated by the new computational challenges posed by the pressing groundwater and energy resources engineering issues to be addressed in the coming decades.The state of the art in modeling groundwater systems is related to developments in geology, physics, and mathematics. However, research in these areas is often pursued completely independently by different teams and is therefore not integrated around a unique vision. Therefore, no advantage is taken of the potential that lies in combining these recent advances. For instance, stochastic geological modeling for alluvial systems using process based knowledge and multiple-point statistics can dramatically increase the ability to generate realistic models of aquifers. However, no geophysical inversion has yet been set up to improve the characterization of such geological structures. Another example is related to computational feasibility. Uncertainty assessment for hydrological systems always necessitates a large amount of simulations based on different setups and scenarios. In many cases, these computational requirements hamper the application of a sound risk analysis. Highly promising solutions to this apparently impossible problem can be found in recent developments in computational statistics. Furthermore, models are always driven by a specific purpose, and exhaustive knowledge of the entire system is usually not required by the modeler. In fact the identification of the relevant processes and complexities are basically unknown unless the entire complexity of the system is simulated. Recent mathematical advances suggest that by simultaneously simulating a limited number of simple and complex models in a joint procedure, the relevant aspects of the models can be identified, and the amount of realizations required for a sound analysis can be dramatically reduced. Like advances in quantitative geology, such promising techniques have not yet been applied in hydrological modeling.The ENSEMBLE project involves 6 research groups covering all major aspects of hydrogeological modeling: geology (University of Basel), geostatistical modeling (Stanford University and University of Neuchâtel), geophysics and fluid mechanics (University of Lausanne), mathematics (University of Bern), and engineering (ETHZ Zürich). Each group is leading one subproject and will closely co-operate with the others towards two highly ambitious goals: 1) Combine the recent but not yet applied advances in each discipline. In order to achieve this goal, all subprojects will share a common numerical framework, common experimental sites, and synthetic data sets for testing the tools and demonstrating the applicability of the integrated method. 2) Beyond combining the already available tools, each group will pursue fundamental research targeted towards the most important factors limiting today’s ability to model hydrogeological systems. These two goals will allow both focused and targeted research.