ABSTRACT. The equations governing free surface flows, typically the Navier-Stokes equations, are difficult to analyse and solve and therefore reduced complexity models are often used to represent geophysical flows. During this presentation, we first present models able to approximate the hydrostatic Navier-Stokes equations. The associated numerical schemes are endowed with stability properties such as positivity, well-balancing and discrete entropy inequality. They are confronted with analytical solutions and experimental measurements. Then we propose non-hydrostatic (dispersive) models and numerical procedures to approximate them. But the numerical analysis of such models is complex and several questions remain opened.

ABSTRACT. We used Hydrogeosphere to investigate how different catchment and event parameters influence transit time distribution (TTD) shapes. If catchments are able to effectively transport incoming flux in the subsurface beyond the outlet, TTDs are gamma functions with high initial peaks. If influx is larger than the outflow capacity, TTDs become humped. The more similar the event inflow is to the outflow capacity and also the larger the available storage becomes, the more the TTDs converge to exponential functions.

ABSTRACT. A new generation of catchment-scale transport models has been developed in the last years where transit time distributions result from the formulation of a general age Master Equation (ME). The new theoretical formulation has improved capabilities, but the numerical implementation of the governing equations is more demanding than in traditional lumped methods as the governing age ME is usually nonlinear. This contribution focuses on the discretization of the age ME and on its solution using general StorAge Selection (SAS) functions.

ABSTRACT. Water age is a useful conceptual framework both for empirically analyzing tracer data, and for translating the resulting understanding into computational tools for water quality prediction. In this talk I will discuss how the development of StorAge Selection (SAS) functions over the last several years has spurred progress toward the goal of improving water quality predictions at the catchment scale, while also clarifying serious challenges that currently hold us back from achieving that goal.

ABSTRACT. The denitrification capacity of an aquifer is determined by interactions of groundwater flow with zones where denitrification occurs. While tools exist to investigate groundwater flow at the catchment scale, predicting the highly spatially variable denitrification zones in aquifers remains challenging. We present a new quantitative framework to infer the location and extent of reactive zones in aquifers from apparent oxygen and nitrate reduction times.

ABSTRACT. Due to poor mixing, fitted reaction rates at field-scale sites are much lower than those found at the laboratory-scale. Pseudomonas Stutzeri strain KC (with sufficient substrate), was injected into aquifer material in the lab and field to degrade carbon tetrachloride (CT) under anaerobic conditions. Eulerian simulations using the ADRE required a 30x decrease in the CT reaction rate upon upscaling. In contrast, the Lagrangian PTR method matches the field-scale experiment using batch-scale rates.

ABSTRACT. To date the Spatial Markov Model has had great success in predicting mean transport - e.g. breakthrough curves, but applications to mixing and reactions are more limited. These are nonlinear in nature and predicting mean behavior is not sufficient. Any model aiming to capture these must account for subscale fluctuations. We propose novel approaches, using up and downscaling methods, to enable this. We start with a simple periodic system and extend to more complex heterogeneous flows.

ABSTRACT. Little is known about the motion of synthetic fibers in water through porous media and there are limited options for direct numerical simulation. This presentation explores particle based methods as a tool for simulating the dynamic motion of synthetic fibers suspended in water. The fibers are represented as freely-jointed chains with attractive and repulsive forces. Motion is simulated as a constrained random-walk added into existing colocation based reaction algorithms, which produces motion that is similar to laboratory observations.

ABSTRACT. One of the approaches for remediating contaminated groundwater in small-scale sites is chemical oxidation. The typical application involves a simultaneous pump-mix-inject approach. Both wells are replaced often to assure a good (chaotic) mixing of reactants. The oxidation of the contaminant is mainly taking place at the aquifer, where mixing is the limiting factor. We present a novel 2D particle tracking model of the process and some preliminary results of the model.

ABSTRACT. The impact of wettability on multiphase flow in porous media continues to challenge our microscopic and macroscopic descriptions. The gap in our understanding could be bridged by pore-scale modeling, which has seen rapid development in recent years and is becoming a useful predictive tool. Here, we cross-validate different pore-scale modeling methodologies by comparing the modeling results from various leading researchers with a benchmark experimental dataset on patterned microfluidic cells (Zhao et al., PNAS 113, 10251–10256 (2016)).

ABSTRACT. Microscale modeling of porous medium systems is an important tool for advancing fundamental mechanistic understanding of flow and transport phenomena. We detail a lattice-Boltzmann method approach to simulate the microscale and compare streaming algorithms, evaluate a new flux boundary condition, present evolving work on adaptive methods, and demonstrate asynchronous multiscale analysis. We also present an analysis of communication and load balancing issues, examine efficiency on modern high-performance computing architectures, and evaluate weak and strong scaling.

ABSTRACT. Dissolutions rates measured at the laboratory differ from those measured at the field scale. There is still no consensus on how to perform the upscaling of dissolution rates, specially when physical and chemical heterogeneities are present. In order to shed light on the source of the scale dependency and to inform the design of an upscaling procedure, we present a methodology to simulate multispecies reactive flow directly on voxelised X-ray micro-CT images of multimineral rocks.

ABSTRACT. A better understanding of sediment erosion and deposition processes is critical to the mission of the US Army Corps of Engineers and many other organizations concerned with flood protection, navigation, and riverine and coastal water resources. In this work we present a computational approach combining finite element and discrete element methods for coupled simulation of air/water/granular flow at the pore scale. The method uses level-set-based immersed and embedded boundary methods for coupling.

ABSTRACT. We present the first 3D hydrogeological model of the entire Lower Yarmouk Gorge (Israel / Jordan) that includes major aquifers, aquicludes, deep-cutting faults and recharge. The model reveals that fault orientation and hydraulic behavior strongly control the location of discharge areas as well as the pressure gradients. The spring temperature is not necessarily linked to thermal convection within faults. Local permeability anisotropy due to aquifers folding or facies changes are features sufficient for the rising of hot fluids.

ABSTRACT. Random Walk Particle Methods(RWPM) are useful for solving heterogeneous advection-diffusion problems without the pitfalls of PDE-based Eulerian methods. This study develops a new RWPM for discontinuous diffusion, based on analytical solutions and the concept of negative mass particles. Previous methods approached the solution by adapting the time step to be smaller near discontinuous interfaces. This novel method can solve transport problems in heterogeneous and discontinuous media with coarse time steps (just a single step in the 1D test case shown)

ABSTRACT. We present two fast methods for the estimation of flow rates in fracture networks. Both are based on the representation of fluid flow in a fracture network as a network graph with nodes and edges. Off-the-shelf graph algorithms are used to obtain an estimate of the flow rate. Comparison of these estimates with direct numerical simulations show that the methods have a relatively high accuracy and very low computational cost, making them excellent for screening purposes and uncertainty quantification.

ABSTRACT. Presented research is aimed at modeling of groundwater flow, governed by Darcy's law, with finite element method on incompatible meshes of combined dimensions. The model is motivated by the reduced dimension approach in which the porous media includes discrete fracture network (typical e.g. for granite rocks). The core of this contribution is coupling 0d-2d and 1d-3d and resolving the singularities which appear in these problems using Extended Finite Element Method.

ABSTRACT. A Farmbot is a small scale fully automated farming system. The FarmBot is a complete open source project providing an easy access to small scale local farming. In this paper, we propose several possible extensions of the Farmbot that raise software engineering challenges, such as the design and integration of different domain-specific languages, the composition of various kinds of models, and the computation of complex relations and correlations from such a set of models.

ABSTRACT. Environmental modelling requires the development of complex computational models, this not only requires the understanding of scientific phenomena but also sufficient experience of operating systems, supporting software and computer architectures. We found environmental modelling, with few exceptions does not fully take advantage of emergent software engineering tools and techniques because tailored tools do not exist. We have identified opportunities for software engineering tools to simplify process and abstract away from the underlying computational complexities, revealing new opportunities for environmental science.

ABSTRACT. The OGC web services standards allow the geoscience community to share data (WMS and WFS) and processes (WPS). Based on the standards, NOUMEA is a Model-Driven Engineering approach to facilitate the development of efficient, reusable and complex WPS workflows. A graphical interface facilitates workflow description and code generation, combining different WPS composition strategies. The framework facilitates the workflow verification, testing and deployment. By hiding technical aspects, the framework allows the scientist to focus on the domain-specific part of the WPS.

ABSTRACT. The Aqui-FR project aims at taking benefits of existing regional groundwater modeling applications, and to develop new products for better water resources management. It currently includes 2 distributed hydrogeological models covering 13 mono or multilayers sedimentary aquifers and a conceptual model covering 25 karstic aquifers in France and more are to be added. We present here the structure of the platform and detail the integration of all the models inside the Open-Palm dynamic coupler along with its future applications.

ABSTRACT. Hyporheic flow in hyporheic systems is highly influenced by in-stream bedforms. An accurate representation of topographical variations of the stream-streambed interface is essential in process-based models in order to represent the couplings between hydrologcial and biogeochemical processes correctly. We present an analytical modeling approach for solving hyporheic flow problems without domain truncation. Applications of the method to hyporheic systems, ranging from the centimeter-scale rippled bedforms to riffle structures of 10m and larger scale, indicate the high accuracy of the approach.

ABSTRACT. To understand the specific nitrate pathways in a catchment, we combine results from longitudinal river water sampling campaigns with small scale groundwater monitoring at the reach-scale. Where groundwater is constantly feeding the river, advective transport of nitrate is the dominant process and denitrification is of minor importance. Increasing 15N isotope values in the stream water from up- to down-stream locations, suggesting an increase in net denitrification activity from the upper to the lower catchment.

ABSTRACT. We developed catchment-scale nitrate budgets for 39 springs in Florida, with springshed areas ranging from 10 to 1000 km2. Nitrate concentrations have been rising in Florida springs over the past 50 years alongside increases in population and changes in land use. Spring discharge integrates the inputs, transport, and attenuation over the entire springshed. Here we evaluate the relative contributions to total N flux that can be attributed to land use, human population density, and soil types associated with N attenuation.

ABSTRACT. The matter of nitrate legacy is of major concern in agricultural, yet there is a lack of quantitative estimations of future nitrate budgets in streams and aquifers. We use a three-dimensional groundwater flow model calibrated with atmospheric tracer data (CFCs) to predict the evolution of nitrate concentrations in an agricultural catchment over the next decades. This prospective approach allows us to identify the key factors controlling aquifer vulnerability to nitrate pollution.

ABSTRACT. In order to protect groundwater, streams, and fjords, the Danish government is currently implementing new strategies based on spatially differentiated application of N fertilizer. Riparian lowlands are acting as important buffers against N seepage to streams and how this takes place is highly dependent on the flow paths. A 3D integrated surface-subsurface model (HydroGeoSphere) was developed for a sub-catchment in Denmark with the major aim to examine the role of the linkage between riparian zones and their upland aquifers.

ABSTRACT. We focus on the longitudinal spread of solute plumes at aquifer scale in terms of mass arrival (the breakthrough curve, BTC) at control planes normal to the mean flow direction. We show by extensive and accurate 3D simulations that the BTCs pertaining to different conductivity structures are quite close, opposite to what happens in 2D flows. We also show that that the bulk of the BTC is predicted quite accurately by the solution based on the First Order Approximation.

ABSTRACT. In the research to be presented, we express concentration or flux of solutes as a distribution over their velocity. We then derive an integrodifferential equation that governs the evolution of the particle distribution over velocity at given times and locations, based on a presumed velocity correlation structure and an ergodic cross-sectional velocity distribution. The transition probability is specified via a copula function that can help construct a joint distribution with a given correlation and given marginal velocities.

ABSTRACT. We study advective transport in heterogeneous porous media. Our analysis show that time or space series of Lagrangian velocities differ. The latter show less intermittency, suggesting a lower correlation. Lagrangian velocity distributions may evolve, depending on the injection mode. We develop two new stochastic models accounting for this property. We compare modeling results to direct numerical simuations; our discussion focuses on the differences between models and their implications for field studies.

ABSTRACT. An accurate modeling of the permeability and diffusion coefficient is essential for predictive flow and transport modeling. Well known models in terms of the porous medium’s porosity are proposed by Kozeny-Carman or Millington-Quirk. Contrarily to these formulae, upscaling methods directly enable to calculate the full, potentially anisotropic, effective permeability or diffusion tensor in terms of porosity or surface area or further geometric quantities. Along this line, we provide new quantitative relations without any artificial fitting parameters.

ABSTRACT. We numerically investigate fluid-fluid displacement in rough-walled fractures with a focus on the combined effect of wettability, viscosity contrast, and fracture surface topography. We develop a model to simulate dynamic displacement of one fluid by another immiscible one. The model is shown to be able to produce compact displacement, capillary fingering, and viscous fingering, as well as the transitions between them. Both reducing the aperture variability and increasing the contact angle (from drainage to weak imbibition) can stabilize the displacement.

ABSTRACT. Motivated by the formation of gas hydrates in seafloor sediments, here we study the volumetric expansion of gas capsules into liquid when the gas-liquid interfaces readily solidify due to hydrate formation. We first motivate the study with a high-pressure microfluidic experiment. We propose a phase-field model that describes the multiphase flow and thermodynamics of the gas-liquid-hydrate system and present high-resolution numerical simulations, which illustrate the emergence of complex crustal fingering patterns from gas expansion dynamics modulated by interfacial hydrate growth.

ABSTRACT. We derive analytical solutions of cyclic CO2 injection into cores to investigate fundamental flow processes related to CO2 huff 'n' puff. Immiscible conditions with incompressible fluids and rock are assumed. The solution is analyzed to study the CO2 distribution, extent of the plume spreading and trapped/immobile CO2. Numerical simulations are conducted to validate the new solutions and also to test their applicability in realistic conditions. The range of parameters in which the analytical solutions apply is delineated.

ABSTRACT. In this study, a proof of concept for a novel kinetic interface sensitive (KIS) tracer is provided through the use of a well-controlled dynamic column experiment in conjunction with a macroscale two phase flow reactive transport model.

ABSTRACT. Non-isothermal immiscible flow of coal tar and water was modeled in a two-dimensional tank to characterize thermal enhancement of coal tar pumping. Experimental setups and numerical simulations were made to better understand the coupling between flow rates and heat transfers. The results show that the pumping of coal tar has a higher recovery rate than pumping at 50 °C than 20 °C. A heating element will be added in the tank in order to study non-isothermal flow.

ABSTRACT. We utilize graphs to increase the computational efficiency of discrete fracture network (DFN) simulations using techniques such as pruning, partitioning and bandwidth reduction. As an alternative to DFNs, we explore graph-based representations of a DFN. We find that the different mappings between the DFN and graph can be used to answer toplogical and hydrological questions. We demonstrate the graphs are an effective tool to increase the computational efficiency of DFNs and in some cases can replace DFN models.

ABSTRACT. Randomly generated fracture networks can display many geometrical difficulties that make the generation of a conforming mesh a very difficult task. In this talk a number of solutions based on an optimization approach to circumvent mesh conformity or based on Virtual Element discretizations to relax conformity constraints are discussed, both: for discrete fracture networks flow and transport simulations and for discrete fracture networks and surrounding rock matrix coupling.

ABSTRACT. A fast upscaling procedure for determining the equivalent hydraulic conductivity of a three dimensional fractured rock is presented in this paper. A modified semi-analytical superposition method is used to account for both the hydraulic conductivity of the porous matrix, of the set of fractures and of the network connectivity. The connectivity indices are determined empirically and new formulation of the superposition method is proposed. The improved model shows good agreement with the numerical results for different configurations of fracture networks.

ABSTRACT. We focus on efficient flow simulations in fractured media following the Discrete Fracture Network (DFN) framework. The complex geometries and the size of the computational domains require an efficient handling of computational resources. Using an optimization based approach, with a standard Finite Element implementation, we can decouple the discrete problem in several local small problems, with few data to be communicated. We propose two parallel approaches, written in C++, one based on the GPU, and one based on the MPI.

ABSTRACT. Direct simulation of solute transport in large DFN's is prohibitively expensive. As an alternative, extracted 3D equivalent pore networks from a dense random distribution of fracture planes are used as computational stencils. Next, a double Laplace transform method is used to solve governing transport equations. It has been shown that this formulation is accurate for any range of local Peclet numbers. Numerical experiments demonstrate the method efficiency to derive equivalent solute dispersion coefficients for large realizations of 3D stochastic fractures.

ABSTRACT. With recent advances in groundbased and airborn electromagnetic methods, groundwater modellers can be equiped unique datasets holding high resolution information on the subsurface structures. These data often reveal a highly complex reality. It is hard to argue that data gaps can be modelled deterministically. Here, we would like to show, how multiple point geostatistical methods can be used in combination with 3D EM data to generate subsurface realizations, and how these realizations can be used efficiently in groundwater models.

ABSTRACT. For modelling river-aquifer interactions, estimating hydraulic conductivity in riverbeds is essential. We suggest using Electrical Resistivity Tomography (ERT) and time-domain Induced Polarization (IP) geoelectrical methods to map the spatial distribution of hydraulic conductivity within riverbeds. The continuous images obtained through ERT and IP are used as training images for multiple-point geostatistical simulations of riverbed hydraulic conductivity. The resulting simulations are used as input to a groundwater flow model to quantify the effect of riverbed heterogeneity on river-aquifer exchange fluxes.

ABSTRACT. Anisotropy is observed in several cases on hydraulic properties in the unconsolidated aquifers in Canada. Geophysics allows linking physical to hydraulic properties so as their anisotropy: to that end, we developed an innovative inversion algorithm to quantify the electrical anisotropy, strongly correlated with the hydraulic anisotropy as we can see on a real case study we present.

ABSTRACT. We discuss a system consisting of a forward and inverse solution for surveys using electromagnetic induction. The forward problem is an improvement of the McNeill model allowing surveys for more conductive soils. Our inverse solution uses regularization which allows piecewise continuous functions. This allows the detection of the interface between two differently conducting mediums (e.g. seawater and groundwater). Furthermore we present an algorithm to determine the regularization parameters during the inversion instead of an a priori value.

ABSTRACT. Geophysical data can be used to constrain geologic uncertainty. Considering a Bayesian approach we use GPR traveltime data to update the distribution of prior uncertainty, i.e. falsify unlikely scenarios. We consider uncertainty in different geologic parameters and model variability with multiple-point geostatistics (MPS). Updating of the prior is done by calculating distances between observed data and forward modeling of the geophysical response of MPS realizations and then projecting in a lower dimensional space where the update is performed.

ABSTRACT. Recent years have seen significant advances in the development and application of morphodynamic models, based on the shallow water equations, to simulate river and floodplain evolution. Despite this progress, significant challenges remain to be overcome if such models are to provide realistic simulations of fluvial system responses to environmental change. This situation reflects a wide range of factors, not least the large number of environmental processes that influence river behaviour, and the fact that these processes operate across a multitude of spatial and temporal scales, many of which cannot be resolved in such models. Here I will discuss some of these challenges by examining the application of morphodynamic models to two contrasting problems. The first application involves the simulation of large sand-bed rivers, which are typically characterized by multiple scales of bed morphology, the finest of which must often be parameterized rather than represented explicitly. High resolution aerial imagery and Digital Elevation Models quantifying dune, bar and channel dynamics in the sandy braided South Saskatchewan River, Canada, will be presented. These data will be used to examine the degree to which depth-averaged morphodynamic models are capable of simulating sand-bed river evolution and responses to high flow events, and to discuss the limitations of existing model process parameterisations. The second application involves the simulation of lowland floodplain construction over periods of centuries to millennia. Hydrodynamic models are regularly used to simulate floodplain inundation and overbank sedimentation during individual floods. However, most existing models of long-term floodplain construction and alluvial architecture do not represent flood hydraulics explicitly, but instead parameterize the effects of hydraulics on processes such as sediment transport and deposition, and river avulsion. Results will be presented from a hydrodynamically-driven model of long-term floodplain evolution, which will be used to explore the controls on channel belt morphology, alluvial ridge construction and avulsion likelihood.

ABSTRACT. Our goal is to numerically approximate hydrostatic free surface flows with variable density. The applications are for instance estuarine water flows, where variations of temperature and salinity result in significant density variations. We start from the Navier-Stokes system with variable density, of which we derive a multilayer discretization. A stable and accurate numerical scheme is formulated and analyzed. 3D numerical simulations are shown. The results are confronted with real data.

ABSTRACT. The derivation of the SW equations is based on depth-integration of the Navier-Stokes equations. We derive a new formulation of the SW equations adapted to a general bottom topography, using an approximation of the ``cross-flow'' surfaces for depth-integration. This yields a set of equations in intrinsic coordinates, characterized by non-autonomous fluxes and sources containing the metric induced by the bottom. We analyze the new SWE and report numerical solution based on a Finite Volume scheme.

ABSTRACT. Rainfall-runoff simulation is growingly being performed with physically-based numerical models. Both shallow-water solvers and zero-inertia approximations to the system have been proposed for this purposes. However, no systematic analysis and comparison of the relative capabilities and performance has been carried out between these approximations in the context of rainfall-runoff simulation. We present a systematic benchmark of these models in this context. Results suggest that both models are applicable, reasonably accurate, and that their computational performance is dramatically different.

ABSTRACT. The “precipiton” method consists of routing elementary water volumes on top of topography with erosive and depositional actions. Here we present an original way to calculate both river depth and velocity from a method that remains embedded in the precipiton framework. We have applied the method to different cases including high-resolution LIDAR topography. When coupled with erosion and sediment transport equations, the model is able to reproduce both straight and braided patterns with geometries independent of grid size.

ABSTRACT. Solving the Shallow Water Equations (SWE) requires the knowledge of river bathymetry, friction parameters, and reach boundary conditions. These detailed properties are not always available, e.g. for discharge-stage modeling at the scale of a full hydrographic network. Here we address this issue at the scale of a geomorphically-significant reach by defining an equivalent periodic geometry, which allows for a spectral solving of the steady-state SWE and the estimation of reach-averaged stage-discharge relations, accounting for high-frequency variations in along-stream river morphology.

ABSTRACT. Flow, sediment transport and ecological processes in saltmarsh environments are important to understand as they are inextricably linked, and exhibit numerous feedback mechanisms. Here, we apply a combined biomechanical-CFD model to simulate flow around flexible saltmarsh vegetation patches. Individual stems are represented by porous blockages within a RANS flow model. Results show good agreement with field data and highlight the importance of flexural rigidity and the vertical shear layer in defining wake length and therefore sediment deposition.

ABSTRACT. CO2 storage combined with enhanced oil recovery (EOR) is perceived as the most cost-effective method of disposing captured CO2 emissions. CO2-hydrocarbon mixtures have non-monotonic density with increasing CO2 concentration that leads to complex convective mixing under gravity-dominated flow. In this paper, we investigate gravity-driven mixing using high-resolution simulation with an accurate property model for CO2 mixtures. The results are important for understanding the impact of CO2 migration in oil reservoirs and more effective storage of CO2 during EOR operations.

ABSTRACT. Mixing in the presence of convective instabilities in porous media is governed by the behavior of stagnation points where the fluid interface is stretched and compressed. We use a interface compression model combined with a stochastic approach to analyze the impact of the conductivity field structure on the fluid interface compression and on fluid mixing behavior. The flow structures are visualized by the strain rate and characterized by their correlation length.

ABSTRACT. Tidal variations can induce complex flows in adjacent aquifers characterized by spatial heterogeneity. Using a simple linear groundwater flow model, we explore the nature of such groundwater flows and show that these can generate chaotic mixing near the tidal boundary. Such "chaotic saddles" in the flow generate non-trivial mixing and transport, and induce significantly augmented reaction dynamics. We uncover the parameters that govern this phenomenon and examine the propensity for such mixing to arise under transient forcing more broadly.

ABSTRACT. In this talk, we present recent experimental evidence of chaotic mixing at the pore scale in 3D porous flows. Based on several measurements of the elongation of a material line, we propose a simple random walk model capable of capturing the main characteristic of chaotic advection in porous flows. We finally discuss the origin of this chaotic behavior and its most significant consequences for upscaling mixing and reactive transport in porous media.

ABSTRACT. We investigated the combined effect of geological structure and engineering-induced oscillating flows on mixing in natural aquifers. By using an analytical framework we analyzed the effect on dilution of alternative oscillating-flow configurations as a function of suitable characteristic parameters, epitomizing aquifer’s heterogeneity and induced-flow patterns. Our results show that inducing forced circulation in the oscillating flow may significantly enhance dilution, while containing the plume. This was not achieved with other flow configurations not including rotational flows.

ABSTRACT. The classical connection between symmetry breaking and the onset of chaos in dynamical systems harks back to the seminal theory of Noether 1918. We study the Lagrangian kinematics of steady 3D Stokes flow through SC and BCC crystalline lattices of close-packed spheres, and uncover an important exception. Whilst breaking of point-group symmetries is a necessary condition for chaotic mixing in both lattices, a further glide symmetry of the BCC lattice generates a transition from globally regular to globally chaotic dynamics.

ABSTRACT. A model of immiscible two-phase flow was developed to simulate gas release from a repository for long lived low and intermediate level nuclear waste and migration to the geosphere. Hydrogen gas is caused by anoxic steel corrosion. Different groundwater flow fields were simulated both at near- and far-field scales. Gas generation does not impact significantly the hydraulic behavior of the system. The hydrogen fluxes to the geosphere are controlled by the gas production rate and the permeability of the host-rock.

ABSTRACT. A three dimensional multicomponent multiphase simulator CubicM is performed combining reactive transport and mass transfer modules. The main purpose is to evaluate and quantify Non Aqueous Phase Liquid plume attenuation by soil microorganisms and kinetic mass transfer processes such as dissolution, volatilization, sorption. The aim of the developed tool is to allow a flexible selection of physical formalisms in order to study the fate of pollutants in time and space. Various test cases are presented with heterogeneous configurations.

ABSTRACT. In many applications, the wettability of the rock surface is assumed to be uniform in time. However, wettability can change permanently when the rock is exposed to an altering fluid. For these systems, the standard capillary-pressure (Pc-S) models are insufficient to describe the physics. Here, we simulate the dynamic Pc-S data by introducing an empirical model for wettability changes at the pore-scale. The resulting Pc-S curves are used to correlate the dynamic term to a Brooks-Corey type power law.

ABSTRACT. Saline formations are attractive for CO2 storage because high salinity renders their associated brines undesirable for future water resources. However, high salinity can lead to dissolved salt precipitating around injection wells resulting in significant loss of injectivity. Mathematical analysis reveals that the process is self-similar and strongly controlled by a dimensionless capillary number. Low injection rates lead to high capillary numbers, which in turn are found to lead to large volume fractions of precipitated salt around the injection well.

ABSTRACT. We present a diffuse interface model of single-component, two-phase flow (a van der Waals fluid) in a porous medium. We propose a Darcy-Korteweg model that is appropriate to describe flow in a micromodel, with a gap-averaged velocity. We study the ability of the proposed model to capture capillary pressure and the dynamics of vaporization/condensation fronts, and show that the model reproduces pressure fluctuations that emerge from abrupt interface displacements (Haines jumps) and from the break-up of wetting films.

ABSTRACT. We undertake a formal derivation of a linear poro-thermo-elastic system within the quasi-static framework. This derivation is based upon the derivation of the quasi-static poroelastic equations from the micro structure, except that we include energy conservation equations in the micro-scale model. These are coupled to the fluid/structure model by using linear thermo-elasticity for the solid instead of the usual linear elasticity. The resulting upscaled system is similar to the linear poro-elastic equations, but with an added conservation of energy equation.

ABSTRACT. We consider two-phase Darcy flows in fractured porous media coupling the d-dimensional flow in the matrix with the (d-1)-dimensional flow in the fractures. Keeping matrix-fracture interface unknowns is a key feature to obtain an accurate reduced model but it also increases the difficulty to solve the linear and nonlinear systems. We investigate in this work the elimination of these interface unknowns at the nonlinear level in order to increase the global efficiency of these models.

ABSTRACT. We consider numerical methods for flow in fractured porous media, with the fractures modeled as lower-dimensional objects that may intersect in complex configurations. We introduce a new framework to handle heterogeneity in terms of both numerical schemes and geometric conformity among the different dimensions. The key ingredient is a conservative, mortar-like approach for inter-dimensional couplings. We present theoretical stability results for the coupling of numerical schemes in the different dimensions, supported by numerical experiments for 2d and 3d fracture networks.

ABSTRACT. Modeling transfers in fractured media remains a challenging task. The major issue to model transfers in fractured media comes from the complex geometry of discrete fracture networks. With the goal to obtain precise flow and transfer simulations we propose a new conforming mesh approach. The idea is to decompose DFN into a number of connected closed outlines. The presentation will explain the meshing procedure and few reference simulations obtained on different fracture networks will also be presented.

ABSTRACT. Aggregation-based model reduction provides a flexible framework for simplifying geometrical complexity associated with fractured porous media. These reduced models are constructed from a high fidelity/high resolution discrete fracture model by aggregating fine grid cells based on geological considerations. The flow characteristics of the reduced model are computed using a flow-based approach. We will show how these techniques can be used to adapt the model resolution to solution requirement to achieve significant model reduction.

ABSTRACT. Convective Darcy-flux approximations are presented for convective fluid transport in porous and fractured media. The methods include Godunov Vt, Va and higher resolution hybrid upwind methods. The convective flux approximations are constructed within a continuous Darcy-flux finite-volume scheme framework. The schemes have control-volume distributed (CVD) flow variables and rock properties, as in standard reservoir simulation, and are comprised of families of multipoint flux approximations (CVD-MPFA), providing consistent convective flux approximations that apply to general tensors on structured and unstructured grids.

ABSTRACT. An algebraic dynamic multilevel multiscale method for multiphase flow in fractured heterogeneous porous media is developed (F-ADM). F-ADM imposes independent multilevel coarse grids on matrix and lower-dimensional discrete fracture domains for all unknowns. It develops an automatic procedure to homogenize or discreetly represent the fractures at coarser levels, by adaptively introducing coarse nodes in the fracture domain. Multiphase 3D cases are studied, where only a small fraction of the fine-scale grid is employed to find accurate complex nonlinear multiphase solutions.

ABSTRACT. Hydraulic tomography (HT) has been shown to be robust in mapping heterogeneity in hydraulic parameters. In this talk, I discuss lessons learned from various laboratory/field studies of HT. In particular, recent research has shown the promising potential for HT to effectively map not only aquifers, but also aquitards. In addition, progress in mapping heterogeneities in fractured rocks will be discussed. Finally, large-scale applications of HT at a pump-and-treat site and a municipal wellfield will be presented.

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. Flow and transport simulations of groundwater are strongly influenced by the contrast and the geometry of subsurface properties. Hydrogeophysical data inversion methods that allow preserving these characteristics exist but often at high computing costs. Here we propose a 3D application of a fast probabilistic inversion algorithm, based on graph cuts and whose efficiency was demonstrated on 2D synthetic cases. It is illustrated by three synthetic cases, to invert oscillatory hydraulic tomography data.

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. A framework for interpreting transient pumping tests in heterogeneous transmissivity fields is developed to infer the overall geostatistical parameters of the medium without reconstructing the specific heterogeneous structure point wise. This method is applied to data of the field site “Horkheimer Insel” (South-West Germany) as well as the aquifer analogon “Herten” to estimate the parameters of heterogeneity from pumping test data of each site.