ABSTRACT. The complex architecture of sedimentary systems produces a range of biogeochemical interfaces within groundwater systems. Alternatively, interfaces can form purely from the imprint of the hydrologic and biogeochemical processes. Both types of interfaces may control water quality by moderating the fluxes of electron acceptors and donors to groundwater. However, the ultimate influence of these ubiquitous interfaces is often difficult to ascertain because of difficulties in measuring gradients in solutes at the appropriate scales. Reactive transport modeling combined with global sensitivity analysis provides an approach to evaluate how such complex interfaces moderate water quality. Examples of high-spatial resolution field measurements combined with reactive transport modeling suggest that both the intrinsic sedimentary architecture and the climate- and human-driven hydrologic forcings combine to create transiently reduced zones (TRZs) that respond over seasonal timescales. The spatial distribution of these dynamic zones not only impacts the persistence of the interfaces but also the degree to which they influence water quality in the advective zones.

ABSTRACT. Quantifying the available freshwater for human use and ecological function depends on fluxes and stores that are hard to observe. Evapotranspiration (ET) is the largest terrestrial flux of water behind precipitation but is observed with low spatial density. Integrated hydrologic models that simulate continental extents at fine spatial resolution are now becoming an additional tool to constrain fluxes and address interconnections. These example applications motivate a path forward to better use simulations to understand water availability.

ABSTRACT. The development and application of an integrated surface and subsurface model for quantifying stream-aquifer interactions along the a large dam-regulated reach of the Columbia River Corridor, as well as their relations to the partitioning of the surface energy budget, will be discussed. Observations from three eddy covariance flux towers along the same reach confirmed the findings from model simulations. On-going efforts toward building a new integrated watershed modeling framework that explicitly accounts for river corridor processes will be reviewed.

ABSTRACT. Strong feedbacks between large systems of shallow water bodies and climate- and human-controlled aquifers are common. Nebraska Sand Hills is grass-stabilized dune region containing thousands of shallow closed-basin lakes and wetlands in hydraulic connection with the Northern High Plains aquifer. Simplified flow model uses groundwater recharge (GR) as the major driver of the system. To constrain uncertainty, we used 48 various projections of GR from GCMs in the 21st century. Results indicate feasibility of the proposed approach.

ABSTRACT. While numerical solutions of groundwater flow could in theory be used to incorporate groundwater dynamics in earth system models, the large spatial and temporal scales involved make these solutions too computationally demanding in practice. We present an approximate analytical solution of unconfined groundwater flow that accounts for three-dimensional dynamics. The accuracy of this solution is analyzed through the comparison with a three-dimensional solution, and its relevance is assessed by comparing measured and modelled water table heights in the United States.

ABSTRACT. Microbially enhanced coal-bed methane production is an innovative idea to stimulate biogenic coal-bed methane production by providing nutrients to the native microbial community. Batch experiments serve as calibration for a numerical model. The model and the reaction kinetics are calibrated and validated for the batch scenario and extended to simulate column studies with reactive transport. The model can be used to test hypotheses that are not easily tested using experiments.

ABSTRACT. we explore on the basis of a pore-scale model approach the link between pore-size heterogeneities and substrate inhibition. Our results show that (1) pore-scale heterogeneities can noticeably promote degradation rates of a self-inhibiting substrate within certain levels of concentration, (2) the effect reverses when the concentrations fall to levels essential for microbial growth, and (3) an engineered combination of homogeneous and heterogeneous media can increase the overall efficiency of bioremediation.

ABSTRACT. To understand the dynamics of a PCE-dechlorinating microbial community, we developed a model to describe organohalide-, iron-, and sulfate-respiring bacteria cultivation in three-phase batch systems. Matlab-IPhreeqc coupling is proposed to simulate interphase mass transfer, microbial respiration, and geochemical reactions. The model was used to interpret experiments investigating the individual/combined impacts of iron- and sulfate-reducing bacteria on PCE dechlorination.

ABSTRACT. We focus on the transport behaviour of Cu under conditions related to a biohydrometallurgical leaching approach using neutrophilic microorganisms in neutral to slightly alkaline solutions. The effect of the DFOB as a model leaching organic ligand on mobility and transport of Cu in the presence of kaolinite was investigated in column experiments. A geochemical transport model was established to describe the transport behaviour of Cu in dependence on geochemical conditions.

ABSTRACT. The U.S. Department of Energy, Geothermal Technologies Office (GTO) is funding a collaborative investigation of enhanced geothermal systems (EGS) processes at the meso-scale. This study, EGS Collab, is a unique opportunity for scientists and engineers to investigate the creation of fracture networks and circulation of fluids across those networks under in-situ stress conditions. This paper describes specific numerical simulations executed to design the experiments, and then provides comparisons of modeling predictions and experimental observations.

ABSTRACT. Predictive subsurface modeling can be used to govern and support re-development decisions for geothermal energy production applications. An assessment of re-development of a low enthalpy geothermal plant by use of reservoir modelling is presented. Two options are examined; either a new well is implemented as an additional injection well to lower injection cost, or a new well is drilled as a replacement well for the existing production well to postpone cold water breakthrough.

ABSTRACT. The characterization of flow and transport in fractured media is particularly challenging. Here, we derive analytical expressions for the retardation of the thermal breakthrough peak amplitude for different fracture geometries that may be very useful to predict thermal transport from solute tracer tests or to characterize the fracture geometry from thermal tracer tests. We use those expressions to interpret the results of thermal tracer tests performed in a crystalline rock aquifer at the experimental site of Ploemeur (H+ observatory network).

ABSTRACT. This contribution presents a multiscale continuum modelling approach for water and nutrient transport in 3D root-soil system. By exploiting the rhizosphere models at root segment scale, the prediction of nutrient uptake rate at root system scale has improved significantly, but the simulation is still at a low computational cost. Under various in silico experiments, the virtual root-soil system has shown its advanced modelling method across various scales and its benefit in studying virtually root-soil interactions processes under different soil conditions.

ABSTRACT. We study the propagation of uncertainty related to hydraulic parameters and ecohydrological fluxes that affect water flow, solute transport, and biogeochemical reactions relative to modeling of dispersion and degradation of glyphosate herbicide and its byproducts. We discuss the features of the probability density functions displayed by modeling predictions obtained by the numerical simulation of the system with stochastic parameters and boundary conditions. To this end, we employ a suite of selected tools of global sensitivity analysis.

ABSTRACT. The filamentous algae Cladophora is one of the most prevalent algae worldwide. It interferes with water supplies, and negatively affects dissolved oxygen and pH in lakes and rivers. Yet, the complexity of its response to temperature, nutrients availability and hydrodynamics challenges our modeling capabilities. We developed a fully coupled computational model able to capture the temporally varying spatial distribution of Cladophora in the Clark Fork river. The model predictions match with remote sensing measurements of the spatial coverage of Cladophora.

ABSTRACT. A significant difficulty in coupled land/atmosphere DA is that the atmospheric boundary layer varies at higher frequencies than the underlying land surface states. We use an EnKF assimilation with a simplified model to explore the assimilation of observations of a high frequency state for updating model estimates of a low frequency state. We then design a coupled land/atmosphere DA that assimilates temporally averaged boundary layer observations and is robust to time scales differences between the land and atmosphere.

ABSTRACT. This study investigates how the assimilation of satellite observations of soil moisture variation (via Brightness Temperature observation) into a large scale conceptual hydro-meteorological model can help in reducing forecast errors and uncertainties. The Murray-Darling Basin is used as a test case and long records (6 years) of SMOS observation are assimilated. The assimilation experiment show promising results with substantially improved soil moisture forecasts.

ABSTRACT. We coupled the parallel data assimilation framework (PDAF) to TerrSysMP, a model for the fully coupled modelling of water and energy transport in the atmospheric, land surface and subsurface compartments. The framework is designed for high resolution models with many unknowns. A series of examples at different scales illustrates the performance of the data assimilation system and the potential of different types of measurement data to improve model predictions.

ABSTRACT. In a climate change context, availability of real time and forecast data is essential for decision makers. Associated with threshold values those data can indicate if a crisis situation is expected (drought or flood). Piezometric raw data from sensors are exposed in interoperable formats and services. This tool is able to cross data from different networks (meteorology, river flow, piezometric) in order to characterize in almost real-time groundwater quantitative state. A web prototype was deployed on 6 case studies.

ABSTRACT. Numerical schemes used to solve time dependent problems can be classified in two groups, attending to the time evaluation of the unknowns: explicit and implicit methods. Conceptual simplicity is the most valuable characteristic of explicit schemes whereas implicit schemes offer unconditional stability. Despite their simplicity, explicit schemes are restricted by stability reasons. However, it is possible to relax the CFL condition (Large Time Step methods, LTS). Both implicit an LTS schemes are tested in this work.

ABSTRACT. The numerical resolution of differential equations or systems in rectangular 2D meshes using techniques of implicit temporal discretization is characterized by a structure of five diagonals within a sparse matrix. In this work, two widely used ways to find the solution of the system of equations are explored and compared in terms of performance and efficiency: Penta-Diagonal Matrix Algorithm and BiConjugate Gradient Stabilized. This comparison is done in three 2D non-linear problems: Burgers' equation, Zero-Inertia model and Shallow Water equations.

ABSTRACT. While finite element analysis is a well-established tool in science and engineering, traditional solver approaches become inefficient or even break down entirely for strongly heterogeneous problems and anisotropies. The topic of this talk is the GenEO spectral coarse space which can be used to either construct preconditioners that perform well in these situations or to generate efficient coarse representations in multiscale applications. A highly scalable implementation in the DUNE numerical framework will be presented alongside numerical results demonstrating its effectiveness.

ABSTRACT. We consider precipitation-dissolution reactions of minerals assuming a variable mineral surface area, coupled to equations of reactive transport. Typical approaches of modeling the mineral surface lead to terms which are not locally Lipschitz continuous. This leads to non-uniqueness of solutions and, related to this, to numerical difficulties. To avoid these difficulties, we propose a transformation which lets the non-Lipschitz term vanish. We also deal with the issue of non-negativity of the solution.

ABSTRACT. The SUNDIALS Suite of Nonlinear Differential-Algebraic Solvers and Integrators includes highly robust and adaptive time integration and nonlinear solver methods and software for ODEs and DAEs. In this talk, we will overview SUNDIALS and present results of the application of some of the packages in SUNDIALS to various systems in water resources applications. Lastly, we present future plans for the SUNDIALS library.

ABSTRACT. Compound-specific stable isotope analysis (CSIA) is increasingly being used for source identification and degradation assessment of organic pollutants in the environment. Reactive transport modelling (RTM) enables simulation of isotope fractionation effects following biological, chemical, and physical processes. RTM thereby serves as framework for quantitative interpretation of CSIA data. This presentation discusses recent RTM work on chlorine and hydrogen isotope fractionation during sequential reductive dechlorination; aquitard-aquifer interaction carbon isotope effects; and carbon isotope fractionation of diffuse pesticide pollution at catchment scale.

ABSTRACT. We present a multi-component reactive transport treatment of rate-dependent isotopic fractionation factors stemming from biologically catalyzed redox reactions. Variable fractionation arises from the balance between three fractionating components of the thermodynamically and isotopically consistent Monod rate expression. This approach supports treatment of isotope fractionation in systems which transition from nutrient rich to nutrient limited conditions.

ABSTRACT. The concept of representative elemental volume has been widely investigated in the context of many primary hydrogeological parameters (e.g. hydraulic conductivity, porosity). To date less consideration has been given to exploring the concept in the context of compound specific isotope analysis (CSIA). This talk will address this issue in the light of recent developments in high resolution monitoring of stable carbon isotope signatures of organic hydrocarbon pollutants in groundwater and novel sampling technologies developed to support this.

ABSTRACT. The reactive transport of dichloromethane (DCM) in porous medium is studied in laboratory aquifers, by coupling genomic, isotopic, hydrological, mass exchange and geochemical characterization. Modelling of associated processes is performed by a two phase and reactive transport model (RTM). The main objectives are (i) to examine how water table fluctuations impact biogeochemical and bacterial community-driven DCM degradation processes and water-to-gas volatilization, and (ii) to define the best-suited formalism to describe these groundwater processes by including isotope and microbial datasets.

ABSTRACT. Bank filtration is an efficient and widely extended technique to increase drinking water production and groundwater reservoirs. Its effectiveness strongly depends on transient behaviour of the hydraulic and meteorological conditions, whose may cause significant alterations on the flow and reactive transport processes in the subsurface. This research attempts to address the identification, simulation, and quantitative evaluation of relevant relations between the emerging contaminant attenuation and the seasonal variations of temperature and redox conditions during a large-term bank filtration operation period.

ABSTRACT. We describe the experience attained during the calibration of a numerical model of saltwater intrusion on physical evidences inferred from a large-scale experiment of laboratory. The physical model represents the seaside part of a costal aquifer and it is built at the University of Padova (Italy) into a flume 500 cm long, 30 cm wide and 60 cm high, filled with glass beads. The numerical model is based on a computational approach which ensures accurate concentration and velocity fields.

ABSTRACT. A sparse channel network and a discrete fracture network with uniform parallel plate fractures were built based on a common structural model for the TRUE – Block Scale site at the Äspö HRL. Both models were manually calibrated against throughflow values from dilution tests in borehole intervals. The two models were later extended to reproduce the cumulative recoveries from short tracer tests with pumped sections along three different flow paths and their predictions for long-term solute transport were compared.

ABSTRACT. CO2 residual trapping is one of the key trapping mechanisms for CO2 geological storage. Yet, very few field experiments have attempted to determine the magnitude of this parameter in-situ. This paper presents the results and model interpretations of two dedicated push-pull CO2 injection experiments carried out at Heletz, Israel, pilot injection site during 2016-2017, specifically tailored to determine the effective in-situ values of CO2 residual trapping.

ABSTRACT. The Perth Basin now has the biggest concentration of geothermal projects in Australia and is one of the main fluvial HSA play in the world. MAR in is also well developed. This study shows that the production and injection bores’ horizontal and vertical layouts with respect to fluvial facies can affect key project design criteria. The study includes a characterisation of the fluvial facies of target aquifers and incorporation of the data into a detailed Feflow finite element numerical model.

ABSTRACT. Urban expansion and associated land clearing drive habitat loss, creating a fragmented landscape where native vegetation is restricted to small natural reserves. Focusing on a small reserve embedded in a suburb of Melbourne (Australia), this study aims to investigate its water dynamics using the physically-based spatially-distributed hydrological model CATchment HYdrology (CATHY). Applying such a modeling approach to small spatial scales permits a more detailed description of the variations of land-surface properties, such as topography, soil heterogeneity, and vegetation.

ABSTRACT. Vegetation self-organisation in water-limited ecosystems is a consequence of scale-dependent feedbacks between water and vegetation. These patterns have been modelled by sets of reaction-diffusion equations representing vegetation dynamics and strongly simplified hydrodynamics. In this work, we propose a ecohydrological model including vegetation dynamics and feedbacks, but with an improved, more physical and rigorous representation of surface hydrodynamics. This allows to simulate new and complex interactions that affect the process of self-organisation and the overall ecohydrological behaviour of these systems.

ABSTRACT. We simulated Alpine ecosystems with an ecohydrological model and related the ecosystem variables to topographic gradients to assess the sensitivity of vegetation to changes in meteorological conditions. This provides an estimation of the vulnerability of Alpine ecosystems to climate change. We found complex patterns of vegetation sensitivity. Inner alpine valleys are currently prone to drought; in a warmer climate, water stress might affect larger Alpine areas. Our results concur with tree-ring and forest inventory data from the Alpine forests.

ABSTRACT. The 2015 El Niño event led to unique drought conditions in the Amazon rainforest. We here show using a combination of solar induced fluorescence and microwave remote sensing that taller and older forests were much less sensitive to droughts compared to younger, shorter, forests using remote sensing observations. Taller forests have deeper rooting depth, are more isohydric and more sensitive to atmospheric dryness rather than rainfall limitations. Forest age and heights are important to simulate rainforest response to droughts.

ABSTRACT. Understanding the emergence of macro-evolutionary phenomena such as stasis and interconnectedness, and how these depend on underlying biotic and abiotic processes forms the key constitutive parts within the larger objective of obtaining a comprehensive macro-evolutionary theory based on first principles.

We will discuss efficient computational approaches within the context of semi-arid ecosystems. Our results show how processes such as speciation and environmental change (both systematic and stochastic) impact the interconnectedness of the system.

ABSTRACT. Hydrologic forecasting is increasingly being recognized as a tool to mitigate impacts of extreme weather. Computationally intensive integrated surface/subsurface hydrologic models can now be employed in next-generation flood forecasting tools. Here we describe a data assimilation and forecasting platform based on the HydroGeoSphere model. Computational challenges for field-data assimilation make traditional data assimilation techniques difficult to implement within large-scale real-time frameworks. To overcome this challenge, we employ methodology based on Ensemble Variational Data Assimilation combined with an evolving initial-conditions library.

ABSTRACT. Calibration of imperfect models might lead to a biased estimation of the model parameters. These biased estimates affects the quality of the model predictions after calibration. We provide a generic framework for accounting for modelling errors during the calibration process as a random variable. The parameters of the modelling error component are jointly calibrated with the imperfect-model parameters. The proposed technique is demonstrated on a number of porous media flow problems simulating multi-phase flow in subsurface reservoirs.

ABSTRACT. Kalman filter based data assimilation algorithms are attractive for inverse modelling of coupled subsurface flow, but their application poses challenges due to Gaussianity and non-linearity assumptions. The computational cost may also be a challenge for large-dimension parameter spaces and “expensive” forward simulators. This study investigates advantages and limitations of Gaussian process emulators in the model forecast, and the parameter sampling techniques that minimize the number of forward runs as well as estimation errors in the GP and its cross-covariance structure.

ABSTRACT. Model predictions of flow in the unsaturated zone require knowledge of the soil hydraulic parameters, which are often unidentifiable and uncertain given the available observations. Data assimilation can be used to account for this parameter uncertainty. In this presentation, we compare different ways of handling parameter uncertainty using the ensemble Kalman filter in terms of soil moisture predictions. We find that the augmented state vector approach is the best approach even when the parameters cannot be identified.

ABSTRACT. Large scale hydrological modeling is limited by the poor knowledge of underground properties. Here we set up a framework to assimilate remote sensing and observation networks to infer parameters of a high resolution hydrological model over West Africa. High compute hours and I/O storage prevent stochastic approaches. We propose the steering of single simulations using an in situ analysis approach (FlowVR) allowing to interact with the dataflow and parameters during runtime in a very modular and portable way.

ABSTRACT. TBA

ABSTRACT. The Eulerian-Lagrangian Localized Adjoint Method (ELLAM) is an interesting alternative to standard methods to solve advection-dispersion transport equation. It preserves the performance of characteristic methods and treats general boundary conditions naturally in their formulations. Several previous studies demonstrated the mass conservation of ELLAM in its formulation and its high computational efficiency for solving advection-dispersion transport equation in a variety of applications. In this work, we present an adaptation of ELLAM to simulate density-driven flow in fractured porous media.

ABSTRACT. We present an approximation of the Shallow Water Equations with friction using continuous Galerkin finite elements and explicit time stepping. We present a method that is second-order in space and third-order in time, positivity preserving, well-balanced with respect to rest states and with respect to steady sliding solutions on inclined planes and robust with respect to dry states. The method uses the Flux Corrected Transport to combine a low-order invariant domain preserving scheme and a high-order method.

ABSTRACT. The computational burden associated with a Monte Carlo-based solution of groundwater flow settings in random transmissivity fields is reduced by using a projection-based surrogate model. The basis functions for the projection are traditionally obtained through the snapshot technique. As an alternative, we propose to use the leading eigenvectors of the covariance computed by solving the equations describing the head moments. This approach enables us to construct a reduced order model circumventing the need to solve the full system model.

ABSTRACT. Simulation of flow through porous media leads to systems of linear equations. For large and ill-conditioned problems, iterative methods together with preconditioning techniques are often used. Recently, the projection space used for the generation of surrogate models, for example, the Proper Orthogonal Decomposition (POD) method, has been combined with preconditioning and deflation techniques to reduce the iterations required by the linear solver. In this work, we study the difference of using the POD-space as deflation-subspace or preconditioner.

ABSTRACT. Tsunami-HySEA is a numerical model to simulate tsunamis generated by earthquakes on multiple GPUs. It is based on non-linear shallow water equations with spherical coordinates. Tsunami-HySEA supports time series, nested meshes with different spatial resolutions, and rectangular and triangular faults to compute the bottom deformation. It has been adopted by several institutions in Europe for their Tsunami Early Warning Systems. Several test problems using realistic topo-bathymetries show the validity and efficiency of the solver.

ABSTRACT. It is a challenge and long-standing problem to numerically describe fluid flows in porous media with permeability variations. Using the traditional numerical scheme, the refinement ratio for the grid cell needs to be increasing dramatically to get an accurate result. We find that this difficulty is caused by the singularities appeared in these problems. The finite analytic numerical schemes are proposed to solve the single-phase and multi-phase fluid flows in heterogeneous porous media to avoid the difficulty of the singularities.

ABSTRACT. A new numerical scheme to simulate fully coupled flow and reactive transport in variably saturated porous media will be presented. The flow is modeled by the Richards equation. In this paper we present a new approach to combine the common linearisation schemes, such as the L-scheme, Picard and Newton methods, in a more efficient way. We will observe an improvement in the computational time, maintaining the accuracy of the classical methods.

ABSTRACT. Recently, multi-phase flow has been on the rise within the reactive transport community. In some applications, the impact of variable porosity is important. An accurate coupling of these processes requires new resolution schemes for the reactive transport algorithms. We present these new approaches for both the transport and the chemistry with the reactive transport code Hytec. In addition to the higher numerical efficiency of these methods, they are applied to the case of the atmospheric carbonation of concrete.

ABSTRACT. Based on the assumption of thermodynamical equilibrium, classical models for two-phase flow in porous media rely on the assumption that the capillary pressure depends only on the saturation of the wetting fluid. However, it is also known that this approach has some shortcomings like hysteresis effects in the capillary pressure curves obtained for drainage and imbibition, as the interfacial area is not considered. Applying homogenization techniques, we investigate the relation between saturation, droplet topology and capillary pressure.

ABSTRACT. This study considers the fate of CO2 injected in deep saline aquifers under supercritical conditions. The Finite Elements Method is coupled with the reduction scheme in order to effectively handle the several equations characterizing such multiphase-multicomponent reactive flows. Several cases are discussed and the influence of different parameters is evaluated.

ABSTRACT. We present a new experimental and numerical investigation of CO2-induced wormhole dissolution in carbonates. Nine new experiments were conducted to quantify the wormhole propagation speed as function of injection rate for 1) single-phase injection; 2) two-phase co-injection; 3) Water Alternating Gas (WAG) cycling. After minor adaptations our previously developed fine-scale and reduced-physics models reproduce the experiments. The main new results are 1) mobile gas phase suppresses wormholing only at lower flow rates; 2) significant wormholing is observed under WAG conditions.

ABSTRACT. Numerical modelling of calcite dissolution by injection of a CO2-rich aqueous solution in an injection well at the 1/20th scale sealed with Portland cement into limestone. The geochemical reactive transport code MARTHE-PHREEQC, using a multilayer radial fine grid with an average cell size of 2.5 mm, has duplicated reasonably well the observed dissolution of calcite near the injection tube, and the formation of wormholes resulting to heterogeneity of hydraulic conductivity.

ABSTRACT. We develop a homotopy continuation method based on dissipation operator for solving multiphase flow and transport in porous media with a fully-implicit scheme. The homotopy is constructed by adding numerical dissipation to the discrete flow equations. Numerical evidences and detailed analysis are provided to illuminate why the dissipation operator can significantly improve the nonlinear convergence of hyperbolic PDE problem. We demonstrate the efficacy of the new nonlinear solver using several challenging examples. The new solver exhibits superior convergence properties.

ABSTRACT. Taking into account water table fluctuations and the interaction of shallow water levels with geothermal systems while modeling their natural state is notoriously difficult. We investigate a new formulation for the non-isothermal compositional gas liquid Darcy flows and its coupling with an advanced soil-atmosphere boundary condition. The model and its formulation are applied on the Bouillante high temperature geothermal field (Guadeloupe, French West Indies).

ABSTRACT. This work discusses a general approach for preconditioning the block Jacobian matrix arising from the discretization of coupled poro-mechanical problems. The basic idea relies on approximately computing an algebraic operator able to decouple the flow and mechanical processes, which can be then solved independently. The decoupling operator is defined by extending the theory of block sparse approximate inverses. The proposed approach is implemented and tested in real-world examples that are used to analyze and discuss the preconditioner properties.

ABSTRACT. Coupled flow and geomechanics models have several applications including groundwater hydrology. Biot's model is often used for describing this coupled behaviour. We discretize the non-linear Biot's model using continuous FE in time and MFEM for the space. The non-linear problem is linearized by L-scheme where the well know splitting scheme fixed stress becomes particular case of our scheme. We consider the case when the volumetric strain and the fluid compressibility are non-linear functions with certain conditions.

ABSTRACT. Iterative solution of linear systems arising from incompressible Navier-Stokes equations is frequently the most expensive computation in hydrodynamic models. Scalable preconditioners have been developed for the fully coupled pressure-velocity problem based on pressure Schur complement and approximate commutator ideas. Unfortunately, these methods perform poorly when naively extended to multiphase flows. In this work we present results on recent variable density and viscosity extensions to two such approximate commutator approaches, the Least Squares Commutator and Pressure Convection-Diffusion preconditioners.

ABSTRACT. We estimated parameters of an empirical and a physical root water uptake model from non-destructive observations of root development, soil water content and potential in two different soils and three water treatments. Root development differed considerably between the different soil and water treatments. These differences could be represented in the physically based model. The inversely estimated parameters of this model compared well with forward calculations that are based on the root architecture and root segment hydraulic parameters.

ABSTRACT. The Branched Transport Problem (BTP) studies optimal resource reallocation in which mass concentration is encouraged, a strategy that leads to the formation of branching structures, ubiquitous in nature. We have recently developed a model coupling an ODE describing transient dynamics for a diffusion coefficient, with an elliptic PDE, able to described the transport infrastructures that resembles the solutions of the BTP. We apply this model to simulate the evolution of plant roots in response to external forcings.

ABSTRACT. Plant leaves largely control land-atmosphere water vapour exchange and are responsible for the vast majority of carbon uptake by vegetation. In this presentation, I will highlight surprising effects of commonly neglected interactions between the leaf energy balance and gas exchange on transpiration, water use efficiency and plant survival, and discuss their potential implications for ecohydrological modelling.

ABSTRACT. Above-ground water storage in vegetation plays an integral role in avoidance of hydraulic impairment to transpiration. We use a novel dataset of biomass water content measured in situ to evaluate the ability of a plant hydrodynamics model, FETCH2, to capture the fast dynamics and diurnal hysteretic cycle of transpiration. The FETCH2 model can reproduce species-specific responses to a broad gradient of meteorolgical conditions and is poised to improve the ability of land-suirface models to simulate transpiration and latent heat flux.

ABSTRACT. Hydraulic performance of plants is governed at the tissue level by physiological traits. At a higher level, the effects of these traits integrate to emergent whole-tree traits. These emergent tree-level traits could be upscaled the canopy and forest scales. Provided a mechanistic description of plant hydrodynamics, these traits could be characterized using the parameters of the formulations describing the water flow through the integrated tree conductive system. We used the FETCH2 model to study tradeoffs and coordination of such traits.

ABSTRACT. It is generally accepted that remote sensing data are very beneficial for flood forecasting. The types of information provided are the catchment soil wetness conditions, and the location and level of floods. This talk presents the use of both data sets for the improvement of flood forecasts. Innovative ways to use soil moisture data for the improvement of modeled hydrographs are presented. Furthermore, a method to estimate channel cross section data for the improvement of modeled flood extents are explained.

ABSTRACT. Localization is crucial for the performance of ensemble-based methods in practical data assimilation problems. Conventional localization techniques typically use the distances between physical locations of model variables and observations to modify the degree of observations’ influence on model updates. They suffer from some difficulties to properly handle, e.g., non-local and 4D observations. In this study, we present an adaptive localization scheme that overcomes these noticed issues in conventional distance-based localization, yet is more convenient to implement and use in practice.

ABSTRACT. Bayesian model selection (BMS) provides an objective guidance of how to select the most appropriate model. Its applicability is however limited by high computational costs. To transfer it to computationally expensive modeling tasks, we combine BMS and model reduction. Accounting for an error in the reduced model, we introduce a novel correction factor in BMS, yielding a model ranking under computational time constraints. We demonstrate our proposed approach on a case study for a stretch of the Rhine river.

ABSTRACT. Our goal is to produce accurate bathymetry estimates with valid uncertainties from video imagery of the nearshore zone. We maintain Kalman filters for wavenumber at selected frequencies, and compute the posterior distribution of depth using the linear dispersion relations and Bayes’ formula. We produce bathymetry estimates with uncertainties over one year at the FRF in Duck, NC and compare the results to monthly ground truth surveys. The results show improved uncertainties over the entire nearshore when compared to cBathy.

ABSTRACT. In iterative ensemble smoother approaches and ensemble methods in general, the ensemble size governs the accuracy of the parameter-estimation procedure. However, consideration of large ensembles may be computationally infeasible in applications with expensive forward solvers. Here, we reduce the computational cost of using large ensembles by employing efficient proxy solvers. To correct the proxy responses for the corresponding model error, the latter of which has the potential to strongly bias posterior ensembles if left untreated, we propose a local-basis approach. In this regard, the discrepancy between the detailed and proxy solver is learned for a subset of the ensemble and stored in a dictionary which grows with each iteration. For each ensemble member, the $K$-nearest neighbours in the dictionary are used to construct an orthonormal basis which allows identification of the model-error component of the residual through projection.

ABSTRACT. This study compares nonlinear input variable selection (IVS) methods against a linear benchmark for synthetic (linear and nonlinear) datasets, a partially-synthetic rainfall-runoff example, and a real-world urban water demand forecasting problem. The results support the general superiority of the nonlinear IVS methods in comparison to the linear benchmark. The differences between the IVS methods are discussed - their strengths and weaknesses highlighted. Future research is directed towards enhancing the different IVS methods to realize overall improvements in forecast accuracy.

ABSTRACT. The huge amount of data from models and observations requires complex integration frameworks, which can deal with high dimensional system state, as well as with various noise characteristics. The present contribution focuses on the development of a data-driven scheme for the cases when the data are subject to non-Gaussian noises. Data-driven approaches have been recently applied in hydrology and suggested promising performances. The proposed method is a hybrid scheme based on a particle filtering requiring a surrogate model for forecasting.

ABSTRACT. In many fields, modelers use multi-model ensemble to combine individual model predictions. Their purpose is to obtain a better prediction than the one of the single best model because one model’s strength could compensate for another one’s weakness. To make the compensation work, the models must not be too similar, e.g. in terms of explanatory variables or underlying assumptions. The poster presents a method to identify redundancy, gaps and data match based on the distances between the models.

ABSTRACT. We herein focus on incorporating heterogeneous reaction (i.e. a Robin boundary condition) into a random walk algorithm (RW) that models the Advection-Diffusion Equation. Previous works derived a first-order-accurate reaction probability p for annihilation of particles that reach the boundary. In this work, we developed a second-order-accurate p. We illustrate the result by two example problems. We observe an improvement in accuracy, which comes at no additional computational cost.

ABSTRACT. In this study, flow in fractured media is simulated via dual porosity appraoches inverted for all their parameters by relying upon a descent method assisted by adjoint state calculations. The adjoint states are either of continous or discrete type and compared to each other in their efficiency to retrieve heterogeneous parameter fields. Both techniques perform similarly, the continuous form keeping however the advantage of being non-intrusive in the code of the forward problem.

ABSTRACT. This work focuses on calibrating seawater intrusion and geothermal hydrogeological models to provide better estimates of hydrogeological parameters. We compare methods to integrate quantitatively electrical resistivity tomography data with existing information to hydrogeological models. Two hydrogeophysical inversion schemes are developed: an uncoupled and a coupled quantitative approach. In each application, the coupled approach significantly prevails over the uncoupled scheme in terms of reliability of the parameter estimates when no model conceptual error exists.

ABSTRACT. The aim of this study is to understand the possible impacts of the presence of impurity gases in the injected CO2 stream during a carbon storage project. A single well CO2 injection-withdrawal test with trace amounts of SO2 and N2 is to be carried out at Heletz pilot site, and be monitored for hydrogeochemical variations field data will be compared to simulations with a reactive transport model, in order to validate the model and calibrate the parameters.

ABSTRACT. The study focus only on the uncertainties stemming from the sampling of calibration data (flow observations), to determine the variability of the hydrological parameter set and their impact on the estimation of the flow indices (low-flows, annual and monthly flows). For comparison, the uncertainties obtained under the same sampling conditions by fitting classical probability distributions were also investigated. These approaches are tested on 644 gauged catchments set on French territory, representative of all hydrological regimes in France.

ABSTRACT. Temporal changes in snowmelt and runoff events, driven by increases in temperature, will have wide ranging impacts on vegetation dynamics that depend on snowmelt-related export of nutrients, temperature-dependent rates of mineralization, and associated changes in carbon and nitrogen pools. This presentation will focus on coupling a comprehensive plant ecosystem model ecosys with a reactive transport simulator ToughReact to explore the hydrological and vegetation connections along a lower montane hillslope in the East River catchment, Colorado.

ABSTRACT. Groundwater flows in fractured crystalline aquifers are dominated by preferential flow pathways, most importantly the interface between the upper weathered layer and the fissured bedrock. The topography of this interface is susceptible to impact recharge dynamics. To test this, an artificial recharge basin was implemented at a well-equipped site. We monitored groundwater levels, and performed subsequent analytical and numerical modelling which accounted for geometry. Results show preferential flow pathways and their geometry (height and lateral extension) strongly control recharge dynamics.

ABSTRACT. We have studied the role of non-reactive minerals on the localization of dissolution processes during CO2 rich brine fluid injection experiments. We used three types of rocks (100% of calcite, 2/3 of calcite and 1/3 of quartz and 2/3 of calcite, 10% of quartz and 20% of clays). We have shown that calcite dissolution is more localized in the limestone sample than in the two others. The homogeneous dissolution in the two other samples is explained by the non-reactive minerals.

ABSTRACT. Modelling of the unsaturated zone is always affected by uncertainty and model error. The model error can be reduced by using integrated hydrological models that comprise neighboring compartments such as groundwater. Besides, data assimilation methods allow for a reduction of model uncertainty and error. In this work, it will be investigated whether data assimilation in a stand-alone unsaturated model can account for neglected fluxes in the groundwater or if it is beneficial using an integrated model containing the groundwater domain.

ABSTRACT. Flow and transport in deep fractured formations are subject to strong channeling effects. This contribution presents applications of a flexible channel network model to simulate flow and transport in deep fractured formations. The emphasis is on the rapid development of alternative models for calibration and/or comparison with field data to support early characterization and site model development and guide further testing and/or management.

ABSTRACT. This research aims at constructing an integrated tool for simulating water flow and reactive solute transport in the subsurface focusing on the water table interface. This is achieved by coupling the existing HYDRUS, MODFLOW, MT3D-USGS and PHREEQC codes and adding functionalities for the transfer of solute concentrations. Information exchange between the saturated and unsaturated zones occurs at each time step. The result is a comprehensive tool for simulating subsurface water flow and reactive transport in an integrated way.

ABSTRACT. We quantified the mixing by the scalar dissipation rate (SDR) in self-affine fractures. The variable-aperture distribution leads to local fluctuation of the temporal evolution of the SDR, whereas the temporal evolution of the SDR in the constant-aperture fractures is smoothly decreasing as a power-law function of time. In the variable-aperture fractures, predicting the global SDR from the longitudinal SDR is inappropriate due to the non-monotonic decrease of the longitudinal concentration second moment, which results in a physically meaningless SDR.

ABSTRACT. Graph-based model reduction techniques for DFNs that use purely topological information can predict first passage times accurately. However, to predict the later stages of the breakthrough curve with any fidelity, we utilize a mapping of the DFN to a graph that allows the graph network to be treated as a network of pipes with edge-weights reflecting hydrological parameters. By actually solving a flow problem on the graph, one can use the calculated fluxes a posteriori to perform model reduction.

ABSTRACT. A process-based, reactive transport model is used to simulate essential biological, and chemical processes involved in the spread of antibiotic resistance in soil. Column experiments are conduced to obtain data, and parameter values on the transport of Antibiotics, Antibiotic Resistant Bacteria and Antibiotic Resistance Genes for calibration and validation. The model is implemented in R using the “rodeo” and “FME” packages for model construction and parameter estimation respectively. Application of the model is aimed at qualitative and quantitative risk assessment.

ABSTRACT. Numerical simulations using digital images of real clay sample obtained by FIB-SEM are performed, in order to better understand the gas transport in saturated low permeable porous media. In our study, we focus on using LBM and SPH with the principal objective to improve the description of gas transport in highly water saturated clays. Various transport characteristics such as capillary pressure, absolute and relative permeabilities are calculated. The obtained numerical results for liquid/gas permeability are then confronted with experimental data.

ABSTRACT. Transport in natural porous media shows non-Fickian characteristics. From the Lagrangian perspective, non-Fickian transport is a result of process memory, e.g., the particles’ correlation of velocity and direction. Our hypothesis is that by realistically mimicking a particles flow path, we inherently represent process memory. We generate realistic particle trajectories by a segment-wise resampling of DNS-based particle trajectories. Altogether, we present a data-driven simulation approach for a 3D, non-Fickian, finite-Péclet transport process capable of scaling transport across arbitrary Péclet regimes.

ABSTRACT. The calibration of groundwater flow models so that it allows a proper risk analysis of a flow process dependent prediction of interest is often not straightforward. Here we leverage the ease of implementation of iterative ensemble smoother algorithms to assimilate nonlinear flow data and address insofar as possible the estimation of uncertainty over the updated parameters. On top of that, strategies to deal with non-multi-Gaussian ensembles are investigated to increase the model plausibility.

ABSTRACT. Decay chain are characterized by the simultaneous production and consumption of intermediate species which challenges the quantitative analysis of stable isotope signatures. Existing analytical approaches to calculate isotope signatures during biodegradation along decay chains are limited in their applicability and alternatively proposed isotopic mass balance approaches are lacking a full validation, yet. This studies presents and verifies a new analytical solution for transport and biodegradation along decay chains and investigates the validity of the isotopic mass balance approaches.

ABSTRACT. The characterization of the hydraulic properties, and connectivity of major fracture zones is essential to model flow and transport in fractured media. In this study the spatial variability of transmissivity and storativity of a fracture network was investigated through field and numerical hydraulic tomography experiments. Several injection tests were conducted at the Grimsel Test Site in Switzerland. A simplified discrete fracture network approach is developed to estimate transmissivity and storativity values of hydraulically active fractures by inverting the pressure responses.

ABSTRACT. Multi-objective genetic algorithm based optimization model “Reservoirs Operation Optimization considering Sediment Evacuation (RESOOSE)” has been used in the study to optimize hydropower and irrigation benefits of cascade reservoirs (Tarbela and Diamer Basha on Indus River). Hydropower and irrigation benefits of Tarbela Reservoir according to existing rule curves are 336 and 1181 Million US$ which have been enhanced by 93% through conjunctive operation with Diamer Basha Reservoir. The optimization of rule curves have further enhanced the both the benefits to 115%.

ABSTRACT. Concentration organization in heterogeneous porous media is key physical factor driving chemical reactivity. At equilibrium, reactivity depends the concentration distribution and its gradient. As high and low values of concentration and concentration gradient don't superpose, a transport equation for the concentration gradients is solved by means of an adapted particle method. Particles dynamically optimize their organization to provide highly accurate concentration gradients. The global strategy combining separate particle methods for the concentration and its gradient gives optimal predictions of eactivity.

ABSTRACT. Wellhead Protection Areas (WHPAs) are frequently delineated assuming steady-state conditions. However, the influence from nature's transient behavior or anthropogenic causes can lead to significant changes in the computed outline. On this study, we propose an engineered pumping management scheme that optimally adapts pumping-injection rates, based on multi-objective optimization concepts, with the goal of addressing robustness against nature transient influence while at the same time considering additional pumping well management issues.

ABSTRACT. A mathematical model and a numerical model for unconfined flow are presented. These are aimed at usage on unstructured 3D polyhedral grids. The performance of Newton and Picard nonlinear solvers in the framework of unconfined flow model is analyzed on model and real test cases.

ABSTRACT. Understanding a contaminant source may help in a better management and risk assessment of a polluted aquifer. In this work, an ensemble smoother with multiple data assimilation (ES_MDA) is employed to simultaneously identify hydraulic conductivities and contaminant source information, including source location, contaminant release time, duration and mass-loading rate, by assimilating all piezometric head and concentration observations, in space and time, at once. The method is demonstrated in a synthetic aquifer.

ABSTRACT. In the context of radioactive waste management, deep geological repository in indurated argillaceous media is considered as a potential solution. Radionuclides transport through argillaceous media is then of concern and in particular via diffusion processes in pore water of micro-fractures of the Excavated Damaged Zone (EDZ) surrounding the storage cells of the repository. In order to simulate diffusion process inside unsaturated micro-fractures network geometry originating from Opalinus clay μ-tomography samples, we chose to use a Lattice Boltzmann Model (LBM).

ABSTRACT. Crop models are established tools to predict yields under variable environmental conditions. Agricultural digitalisation produces an ever-increasing stream of data that may be used to improve crop management at the farm level or governance at the societal level (e,g,. hazard management). The field-scale data stem from precision farming machinery and sensors. Regional coverage can be obtained by fusion with satellite data products. We applied data assimilation methods to improve yield prediction at different spatial scales and reduce the related uncertainties.

ABSTRACT. Saturated flow and tracer advection/diffusion transport is simulated at the pore scale in the 3D geometry obtained from X-ray CT scans of Fontainebleau sand samples. The analysis focuses on the effect of scan resolution on the value of effective properties (permeability, effective diffusivity and dispersivity). A criterion based on the morphology is proposed to assess the quality of the scan with respect to the computed properties. These numerical predictions are compared with results obtained from column experiments breakthrough curve analysis.

ABSTRACT. Accurate modeling of multiphase flow in hydraulically fractured shale rocks is critical for unconventional hydrocarbon recovery analysis. Towards this effort we develop a high-fidelity discrete fracture model (DFM) to simulate multiphase flow in fractured rocks. The developed DFM model can not only accurately simulate fracture-matrix interactions and their effects on multiphase flow processes in low-permeability formations, but also provide a useful interface to couple with the hydro-fracture model for reliable prediction and assessment of hydrocarbon production and hydraulic fracturing performance.

ABSTRACT. We investigate the impact of heterogeneity on anomalous transport in porous media. To this end, we derive a continuous time random walk model that accounts for different injection conditions and that is parameterized in terms of flow and medium properties only. This model is applied to experimental data to reproduce solute breakthrough curves. We show how to include diffusive processes to the model and we discuss their impact on average transport.

ABSTRACT. Field observational data indicate that some aquatic geochemical components, such as sulfate and calcium, exhibit distinct characteristics under snowmelt and baseflow conditions. Subsurface geologic structure and mineral composition may have a strong influence on C-Q relationships and geochemical responses with different water infiltration and groundwater level scenarios. In this study, a high-resolution reactive transport model (RTM) is developed to quantify the surface and subsurface interactions of hydrology and geochemical processes, including calcium dissolution and pyrite oxidation, in Copper Creek, Colorado.

ABSTRACT. DarcyTools is a multi-physics toolbox developed by the Swedish Nuclear Fuel and Waste Management Company (SKB) for groundwater flow and transport modelling. An unstructured Cartesian grid and a finite volume approach are the key elements in the discretization of the basic coupled equations such as solute mass conservation, thermo-hydraulics, hydro-mechanics and hydro-chemistry equations. A parallel multigrid preconditioned block-GMRES solver (MIGAL) is involved in modified Picard iteration schemes. The importance of linearization techniques and experiences on hybrid parallelization will be discussed.

ABSTRACT. In this work, we investigate the possibility to use sinusoidal pumping tests in unconfined aquifers to quantify the hydraulic properties. Real sinusoidal pumping field tests were performed in an experimental site located in a porous and fractured chalk aquifer. The sinusoidal pumping field tests are simulated using 3D modelling and sensitivity analysis of the variables as a function of the hydraulic parameters is performed. The uncertainty related to the inversion of the hydraulic parameters is examined.

ABSTRACT. Using the simultaneous injection of conservative and reactive tracers combined to a continuous field monitoring of the aquifer biogeochemistry, this study proposes a new field experiment for the characterisation of the physical, chemical and biological controls on solute transport in heterogeneous porous media. This approach is based on a single injection of dissolved noble gases and nitrates and a continuous monitoring of the dissolved gases, anions, cations, trace elements, stable isotopes and the microorganism community.

ABSTRACT. Riverbank filtration is used to improve the quality of the water before injecting it into a confined aquifer for storage and recovery in Busan, South Korea. However, unexpected high concentrations of iron and sulfate have been measured in the riverbank-filtered water. A reactive transport model was developed to test different geochemical processes on the basis of model performance against field data. The results demonstrate the capability of the model to capture changes in water quality and to test different processes.

ABSTRACT. We present a RTM workflow which is based on the direct experimental derivation of effective volume and velocity vector field from sequential positron emission tomograms (PET) of the tracer transport in geologic materials. The focus on tracer concentration incorporates unconstrained upscaling from molecular sensitivity to the millimeter scale. This workflow allows minimizing deviations from the experiment and computational expenses.

ABSTRACT. We summarize the results obtained from a code-inter-comparison study initiated within the TRUST project financed by the European Community’s 7th Framework Programme, addressing the injection of carbon dioxide (CO2) in saline geological formations. Four numerical simulators are used to model multi-phase flow in porous media, and the effects of interests arising during CO2 injection (e.g., CO2 plume shape, fluid pressure and temperature evolution, deformation, etc.) are compared.

ABSTRACT. In ungauged basins, where no streamflow data is available, the parameters of the rainfall-runoff model needs to be estimated by regionalization. Three methods can be distinguished: (i) transposition, (ii) prescription and (iii) constraint. This work presents a combination of these for a spatially distributed rainfall-runoff model. The 12 model parameters are spatialised according to several patterns, from uniform to mesh scale, thanks to the three regionalization methods.

ABSTRACT. In order to quantify the extent of diffusive deuterium isotope fractionation of benzene and toluene, experiments were conducted based on 1-D gel dissection tubes, as well as on a quasi-2-D flow-through setup, using mixtures of nondeuterated and perdeuterated compounds. The results show an unexpected contrasting deuterium isotope fractionation behavior between toluene (normal) and benzene (inverse).

ABSTRACT. The development of geothermal resources requires the understanding of the distribution of temperature. Predicting the temperature cannot be done at the operation scale but must take into account basin scale processes. In fact, simulated diffusive and advective palaeoclimatic phenomena explain the thermal regime of the Anglo-Paris basin. Mixed-convective models clearly show that vertical flow through formations in fractured areas explains thermal anomalies. As different formations in the basin are exploited for their resources, this calls for considering conflicts of use.

ABSTRACT. Modeling of microbially induced calcium carbonate precipitation is used to design and evaluate a field applictation sealing a high-permeable fractured layer in a sandstone formation. Different injection strategies are modeled before the application as well as afterwards the actually used injection strategy. Both field measurements and the modeling results show decreasing formation injectivity, but the model results suggest that the precipitation efficiency of injected calcium in the vicinity of the injection well was low.

ABSTRACT. First, we will present the software PoreChem, used for the simulations and calculations for this study. Then, we give an overview over the topic of homogenization of reactive transport, using the method of asymptotic expansion. Further, we show recent results on pores-scale simulations of reactive transport and the solution of cell problems for homogenizing passive and reactive transport. Finally, we show an application to reactive transport of MCPA in soil and we give computational results and discuss comparisons with experiments.

ABSTRACT. Flux-continuous finite-volume schemes comprised of control-volume distributed multipoint flux approximations (CVD-MPFA) are presented for flow in porous media with fractures. Both cell-centred and cell-vertex approximations are considered and maintain a single degree of freedom per control-volume, per flow variable and provide consistent flux approximations for general tensors on structured and unstructured grids. Development of surface CVD-MPFA approximation and fracture model approximations are discussed. The schemes are applied to problems including fractured anisotropic media, and comparison of scheme performance is presented.

ABSTRACT. The characterization of karst hydrosystems is generaly based on the monitoring of the outlets which are supposed to integrate the whole behavior of the basin. This methodology is not adapted to the case of hydrogeological sub-basins, located at the head of the watershed. A multidisciplinary analysis based on optical imaging, statistical analysis of piezometric data and hydrogeophysic sounding provides critical information for geometric modeling of geological units, fracturated / karstified zone and for conceptual modeling of hydrodynamic site model.

ABSTRACT. Fault zone have a high potential for geothermal energy extraction. But their usability is highly conditioned by their productivity potential which must be estimated prior to drilling. The hydraulic behavior of the reservoir can be investigated through numerical simulations based on conceptual models of fault zone. DFN depicting the internal fault zone architecture are built and flow simulations are then carried out. Results from different architectures are compared to estimate the impact of fault zone architecture on geothermal reservoirs operations.

ABSTRACT. The groundwater flow in the near-field of an underground nuclear waste repository located within a fractured rock is evaluated by different conceptual and numerical approaches. The fractured system is represented by 1) an Equivalent Continuous Porous Media (ECPM) modelled with Darcy Tools and 2) a Discrete Fracture Network (DFN) modelled with COMSOL Multiphysics. The results of both approaches are compared for water flow and solute transport for different model resolutions and complexities.

ABSTRACT. MT3DMS has long been a popular model in the groundwater field. The MOC, MMOC, and HMOC regard the Cauchy boundary as a source condition. For the source, the MOC, MMOC, and HMOC calculate the Lagrangian concentration by setting it equal to the cell concentration at an old time level. However, above calculation is a petty tricky method. To circumvent this problem, a new scheme is proposed, which combines the numerical formulations of FEM and ELM into one global matrix equation.

ABSTRACT. In this talk we present recent results that are relevant for the application of so-called level set methods to track interfaces and boundaries that can occur in computational models used in water resources research. We are especially interested in robust numerical methods that are unconditionally stable with respect to the choice of time steps.

ABSTRACT. We consider Proper Orthogonal Decomposition (POD)-based Non-Intrusive Reduced Order Models (NIROMs) where a Radial Basis Function (RBF) interpolation method is employed as the black-box interpolation scheme, thus avoiding modification of the original high-fidelity code. We compare the performance of the POD-NIROM method with traditional Galerkin and Petrov-Galerkin projection based hyper-reduced POD methods like POD-DEIM and gappy POD. We evaluate the schemes' performance by considering their accuracy, robustness, and speed for test problems representative of dam-break and riverine flows.

ABSTRACT. Injection of CO2 into the subsurface at industrial-scale can result in significant pressure increases in the reservoir that if not properly controlled, can lead to potential environmental impacts such as fault activation, leakage through abandoned wells, or caprock fracturing. This work presents an integrated adaptive management approach involving monitoring, model calibration, and optimization of brine extraction for pressure control. We investigate the optimization performance affected by initial site characterization and dynamic model updates with new acquired data during the injection.

ABSTRACT. Physically-based models represent detailed surface/subsurface transfer, but the required spatial information doesn't allow their operational use. In situ data on pesticides in a catchment are usually rare and not continuous in time and space. Satellite images, on the other hand, well describe data in space, but only water related, and at limited time frequency. This study exploits these 3 types of information (model, in situ data, images) with data assimilation, in order to improve modeling of pesticide and hydrological transfers.

ABSTRACT. A methodology based on micro-CT images is developed to perform direct numerical simulation (DNS) of flow&transport in micro- and macro-pores. Differential imaging provides representation of sub-resolution and macro pores, while DNS incorporates: (i) flow simulated by Navier-Stokes equations, (ii) streamline-based method for advection, (iii) diffusion superimposed by random walk. Exemplar carbonate rocks with high permeability (Ketton limestone), intermediate permeability (Estaillades limestone), and low permeability (Portland limestone) are shown to quantify the impact of velocity distributions on macroscopic flow and transport.

ABSTRACT. We propose a model for transport and reactive flow in porous media. The transport model is based on a Particle Tracking Method based on Continuous Time Random Walks and modelled accordingly to Rhodes and Blunt (2006), with an extension o enable reactive flow predictions. We validate it using Nuclear Magnetic Resonance experiments and capture the time evolution in a reactive flow experiment. Next we analyse the transport evolving non-Fickian signatures during reactive flow and observed an increase in transport heterogeneity.

ABSTRACT. Reduced-order models (ROMs) for reactive mixing in a vortex-based velocity field are developed using machine learning algorithms. Datasets based on high-fidelity simulations of anisotropic reaction-dispersion are used for training the algorithms. A recently developed non-negative finite element formulation is used for the high- fidelity simulations. We show that the developed ROMs are accurate compared to high-fidelity simulations with a R2-score is greater than 0.9 while the computational time of ROMs are five orders of magnitude faster than high-fidelity simulations.

ABSTRACT. In Lagrangian approaches to reactive transport modeling, reactions are driven by the interaction between numerical particles that represent either a volume of fluid or a mass of solute. To compute this interaction, kernel functions can be used. Because of the complexity of solute concentration distributions in heterogeneous media, the globally optimal kernel is often far from being locally optimal. We present an adaptive locally optimal kernel function whose size, shape and orientation is not only time-dependent but also space-dependent.

ABSTRACT. Modeling solute transport in high heterogeneous porous media faces two challenges. First, the dispersion dependency with scale has to be reproduced. Second, a good representation of the mix process is of paramount importance because it leads to chemical reactions. Finding an equation that satisfies these conditions is needed. In this contribution a new formulation to reproduce the solute evolution is proposed. The velocity distribution is discretized as well as space and time. Therefore concentration is also a variable in velocity

ABSTRACT. The CSMP++GEM coupled code allows reactive transport simulations of two-phase flow with boiling and partitioning of chemical species between the phases, and fluid-rock interaction including non-ideal mineral solid solutions. CSMP++GEM uses the control volume finite element method (CVFEM) implemented in CSMP++ for flow, enthalpy-based heat transport and solute transport computations, coupled with the Gibbs energy minimization method from GEMS3K library for solving chemical equilibria using the Sequential Non-Iterative Approach (SNIA). Our reactive transport code was benchmarked against TOUGHREACT and OpenGeoSys-GEM.

ABSTRACT. To evaluate information derived from overdamped slug tests conducted in fractured sedimentary rocks, the effect of wellbore storage is removed by using the deconvolution technique in order to use diagnostic plots. Using a recently available algorithm for computation of the log-derivative for noisy data, this study illustrates that the deconvolution technique can also be applied to short duration tests. In most cases, the flow pattern suggests that slug tests are sensitive enough to heterogeneities.

ABSTRACT. The parameter estimate and linearized uncertainty quantification produced by state-of-the-art geostatistical inversion methods are often based on assumptions that don't hold in practice, like Gaussianity of the posterior distribution.

We present orthonormal residuals, obtained as byproducts of linearized UQ through randomized SVD of the linearized forward model in dimensionless formulation, as an efficient technique for validation of the resulting parameterized models. The applicability of the method is demonstrated through applications implemented in the high-performance scientific computing suite DUNE.

ABSTRACT. Modelling of root-soil interactions has become an essential tool for studying physical biochemical processes in the terrestrial system. However, the soil characteristics and root properties such as hydraulic parameters, nutrient buffer capacity in the soil are mostly uncertain due to lack of our knowledge. This contribution presents an uncertainty quantification study for nutrient uptake model of 3D architectural root system in a 3D soil using Monte Carlo (MC) method, Taylor series approximation (TSA) and Polynomial Chaos Expansion (PCE).

ABSTRACT. We develop of a novel method for a continuous measurement of microbial activity in groundwater. The method is based on the use of Fluorescein DiAcetate (FDA) whose product of reaction can be measured continuously by a field fluorimeter. Tested and validated in the lab, the method was used in the field during reactive tracer test experiments. A kinetic model allowed us to interpret our results. This method thus opens new perspectives for the characterization of biogeochemical processes in the field.

ABSTRACT. Hyporheic zones can serve many roles throughout the year and may provide a critical control point on watershed scale exports. We apply loose numerical coupling to represent groundwater-surface water connectivity, biogeochemical cycling, and microbial growth in a hyporheic zone test case. Microbial growth is included in the numerical code MIN3P to represent the physically based feedback when pore-spaces are “bioclogged”. A Bayesian MCMC approach is used to constrain the river GPP model to data collected at the East River, Colorado.

ABSTRACT. We present a detailed investigation of electrochemical effects during conservative and reactive multicomponent ionic transport, under advection-dominated conditions, in homogeneous and heterogeneous domains. Laboratory experiments were performed in a quasi 2-D flow-through setup by using pulse injection of electrolytes. We periodically measured ions’ concentrations, at different spatial locations at the outlet, revealing remarkable differences in breakthrough and mixing behavior. We also show numerical simulations, with a recently developed 2-D code, showing the important influence of electrostatic effects during reactive transport.

ABSTRACT. We develop an a posteriori-steered algorithm for the two-phase compositional flow with exchange of components between the phases in porous media. The discretization relies on the backward Euler scheme in time and the finite volume scheme in space. The resulting nonlinear system is solved via an inexact semismooth Newton method treating the phase transition. Numerical experiments are given for the semismooth Newton-min algorithm and the GMRES solver, showing good quality of the estimates and of the adaptive stopping criteria.