Understanding the retention of per- and polyfluoroalkyl substances (PFAS) in the vadose zone is vital to the management of impacted sites. This paper examines PFAS retention in the unsaturated zone above the water table using a mathematical model, MODFLOW-USG-Transport PFAS or "USGT-PFAS." The USGT-PFAS model incorporates adsorption onto air-water interfaces, providing a more comprehensive understanding of PFAS retention near the water table and release to groundwater. Modeling of a hypothetical perfluorooctane sulfonic acid (PFOS) site under various idealized site conditions illustrated that the impacts on PFOS retention from smallest to largest were water table fluctuations, low episodic recharge, constant recharge, moderate episodic recharge, constant recharge with water table fluctuations, and high episodic recharge. PFOS retention also varied by sand type, with greater retention occurring in simulations incorporating coarse sand with low capillary potential versus fine sand with high capillary potential. PFAS management strategies were also explored, including the adaptation of gas sparging, a method traditionally used for volatile organic compounds. Gas sparging can concentrate PFAS in groundwater and the vadose zone around the water table, facilitating retention or removal. Model simulations for a simplified hypothetical site demonstrated that PFAS can be substantially retained in the unsaturated zone once gas sparging results in an upward concentration of PFAS in groundwater and the unsaturated zone near the water table. Modeling can aid in understanding PFAS behavior but requires simulation of multiple interrelated processes to correctly predict PFAS fate and transport in subsurface conditions.

Managed aquifer recharge has become a standard water resources management practice to promote the development of locally sustainable water supplies and combat water scarcity. However, installation of injection wells for replenishment purposes in urban areas with complex hydrogeology faces many challenges, such as limited land availability, potential impacts on municipal production wells and known subsurface contamination plumes, and complex spatially variable hydraulic connections between aquifer units. To assess the feasibility and cost-effectiveness of injecting advanced treated water (ATW) into a complex urban aquifer system, a Simulation-Optimization (SO) model was developed to automate a systematic search for the most cost-effective locations to install new wells for injecting various quantities of ATW, if feasible. The generalized workflow presented here uses an existing MODFLOW groundwater model-along with advanced optimization routines that are publicly available-to flexibly accommodate a multiobjective function, complex constraints, and specific project requirements. The model successfully placed wells for injection of 1 to 4 MGD of ATW in aquifers underlying the study area. The injection well placement was primarily constrained by avoiding excessive impact on environmental sites with underlying groundwater plumes. The largest costs were for well installation and piping to the wells from the existing ATW pipes. This workflow is readily adaptable to other sites with different complexities, decision variables, or constraints.

The performance of time series models is assessed using synthetic head series simulated with a numerical model that solves Richards' equation for variably saturated flow. Heads were simulated in a homogeneous unconfined aquifer between two parallel canals; measured daily precipitation and potential evaporation are specified at the land surface and root water uptake is simulated. The head response to a precipitation event is nonlinear and depends on the saturation degree and rainfall before and after the precipitation event while evaporation reduction occurs during summers. Synthetic series were generated for 27 years and three different soil types; the unsaturated zone thickness varies between 0 and >5 m. The synthetic head series were simulated with a linear and nonlinear time series model. Performance of a linear time series model with four parameters, using a scaled Gamma response, gave R-2 values ranging from 0.67 to 0.96. The nonlinear time series model with five parameters simulates recharge using a root zone reservoir after which the head response to recharge is simulated with a scaled Gamma response function. The nonlinear time series model was able to simulate all synthetic head series very well with R-2 values above 0.9 for almost all models. The head response of the nonlinear model to a single precipitation event compares well to the response of the variably saturated groundwater model. The provided scripts may be used to simulate synthetic head series for other climates or for systems with additional complexity to assess the performance of other data-driven models.

MODFLOW 6 is the latest in a line of six ''core'' versions of MODFLOW released by the U.S. Geological Survey. The MODFLOW 6 architecture supports incorporation of additional hydrologic processes, in addition to groundwater flow, and allows interaction between processes. The architecture supports multiple model instances and multiple types of models within a single simulation, a flexible approach to formulating and solving the equations that represent hydrologic processes, and recent advances in interoperability, which allow MODFLOW to be accessed and controlled by external programs. The present version of MODFLOW 6 consolidates popular capabilities available in MODFLOW variants, such as the unstructured grid support in MODFLOW-USG, the Newton-Raphson formulation in MODFLOW-NWT, and the support for partitioned stress boundaries in MODFLOW-CDSS. The flexible multi-model capability allows users to configure MODFLOW 6 simulations to represent the local-grid refinement (LGR) capabilities available in MODFLOW-LGR, the multi-species transport capabilities in MT3DMS, and the coupled variable-density capabilities available in SEAWAT. This paper provides a new, holistic and integrated overview of simulation capabilities made possible by the MODFLOW 6 architecture, and describes how ongoing and future development can take advantage of the program architecture to integrate new capabilities in a way that is minimally invasive and automatically compatible with the existing MODFLOW 6 code.

Simulating the interaction of groundwater with surface water networks using traditional boundary packages available with MODFLOW-USG can be challenging for complex systems. Often several package types are required as they are typically purpose built. Moreover, these packages generally do not interact with one another which complicates accounting of groundwater discharge at different points within the system. Here, we demonstrate that the connected linear network (CLN) package of MODFLOW-USG, and advances therein in USG-Transport, can be used to simulate groundwater interaction with a complex surface water network comprised of creeks, ponds, wetlands, and springs, in a manner that is comparable with these other packages, but with additional benefits, including explicit routing of water between the features.

First posted March 3, 2022 For additional information, contact: Director, Integrated Modeling and Prediction DivisionWater Mission AreaU.S. Geological Survey12201 Sunrise Valley Dr., MS 411Reston, VA 20192-0002Contact Pubs Warehouse This report documents a new Groundwater Transport (GWT) Model for MODFLOW 6. The GWT Model simulates three-dimensional transport of a single chemical species in fowing groundwater based on a generalized control-volume fnite-difference approach. Although each GWT Model is only able to represent a single chemical species, multiple GWT Models may be invoked within a single MODFLOW 6 simulation to represent solute transport of multiple non-interacting chemical species. The GWT Model is designed to work with the Groundwater Flow (GWF) Model for MODFLOW 6, which simulates transient, three-dimensional groundwater fow. The version of the GWT model documented here must use the same spatial discretization used by the GWF Model; however, that spatial discretization can be represented by regular MODFLOW grids consisting of layers, rows, and columns, or by more general unstructured grids. The GWT Model simulates (1) advective transport, (2) the combined hydrodynamic dispersion processes of velocity-dependent mechanical dispersion and molecular diffusion, (3) adsorption and absorption (collectively referred to as sorption) of solutes by the aquifer matrix, (4) transfer between the mobile domain and one or more immobile domains, (5) frst-or zero-order solute decay or production, (6) mixing from groundwater sources and sinks, and (7) direct addition of solute mass. The GWT Model can also represent advective solute transport through advanced package features, such as streams, lakes, multi-aquifer wells, and the unsaturated zone. If the GWF Model application uses the Water Mover (MVR) Package to connect fow packages, then solute transport between these packages can also be represented. The transport processes described in this report have been implemented in a fully implicit manner and are solved in a system of equations using iterative numerical methods. The present version of the GWT Model for MODFLOW 6 does not have an option to calculate steady-state transport solutions; if a steady-state solution is required, then transient evolution of the solute must be represented using multiple time steps until no further changes in solute concentrations are detected.

In recent years, among renewable energies, the geothermal resource exploitation shows a constant growth; specifically, in countries engaged in CO2 emissions reduction and dependent on energy from abroad, the low -temperature geothermal energy (geo-exchange) for air conditioning of buildings represents a cost-effective and green solution. In closed-loop systems borehole heat exchangers (BHE) are coupled with ground-source heat pumps (GSHP) constituting the key component of the heating ventilation air-conditioning (HVAC) sys-tem. Therefore, the design of the BHE and the correct interpretation of in situ Thermal Response Tests (TRT) are essential to supply the building energy demand. To support the design, Modflow-USG Connected Linear Network (CLN) and Drain Return Flow (DRT) packages are adapted and improved to reproduce the operation of one or more BHE in aquifers and to analyze the TRT. The improvements are compared with a previously developed numerical model and two different analytical solutions (infinite line source and moving line source) by imposing a constant heat rate injection into the aquifer. The results show good agreement between the new approach and previous ones (discrepancy lower than 2% for models with highly refined grid), but the new approach is much more accurate and expeditious in both implementation and execution, also allowing for an easy numerical simulation of multiple BHE.

The χMD matrix solver package is incorporated into USGS groundwater modeling software, such as MODFLOW‐NWT, MODFLOW‐USG, and MT3D. The solver is used to solve matrices assembled through numerical discretization of the groundwater flow equation, and solute transport equations. χMD has demonstrated its higher robustness, faster execution speed, and more efficient memory usage compared to the existing solvers for many types of groundwater flow problems. χMD uses preconditioned iterative Krylov‐subspace methods and consists of preconditioning and acceleration modules. Because the solver package uses a variety of preconditioning features including level‐based incomplete lower‐upper (ILU) factorization method with a drop tolerance scheme, users must choose optimal preconditioning parameters to improve execution speed and robustness. In order to examine how the preconditioning parameters, ILU factorization level, and drop tolerance values affect the overall performance of the matrix solver, we evaluated five different groundwater model applications using MODFLOW‐USG that include different numerical complexities. For those five cases, the number of discretization nodes varied from 10,000 cells to 730,300 cells. From the analysis, we found that the preconditioning parameters greatly affect execution times and memory usage of the preconditioning and acceleration procedures. In addition, a combination of the ILU level between five to seven and the drop tolerance value between 10−2 and 10−3 usually resulted in shorter overall execution time. Our study suggests that the users can elicit higher performance and robustness of the χMD matrix solver using this combination of the parameters and enhance computational efficiency of solving groundwater and solute transport problems.

Simulating heat transfer in an aquifer with one or more vertical Borehole Heat Exchangers (BHEs) of a Ground Source Heat Pump (GSHP) system by means of a finite difference code is difficult because of the square or rectangular geometry grid and computational times thus limiting the types of evaluations that can be performed. The aim of this work is to explore through MODFLOW-USG code (public domain software) a different approach towards simulating a borefield that would be more efficient computationally, in order to enable simulations of larger domains with multiple BHEs. The Connected Linear Network (CLN) package, introduced in MODFLOW-USG, generally simulates 1-D linear computational cells in a 3-D grid, such as hydraulic pipes in subsoil, but for the first time has been adapted to reproduce vertical closed loop U-pipe of a BHE. Therefore, this work evaluates the MODFLOW-USG and CLN package capability to reproduce the yearly operation of one or more BHEs in an aquifer as a simpler and faster approach compared to a very fine finite-difference discretization. Once the CLN package was adapted, a sensitivity analysis on the grid size refinement was performed. There were several findings from this work. The results of the different numerical models were in good agreement with an already validated model, in terms of exchanged energies and aquifer thermal perturbation. Same analyses were carried out for different groundwater flow velocities and it was confirmed that the exchanged energy by a BHE increases with the groundwater flow velocity in accordance with literature studies. At last, a borefield of 7 BHEs was implemented in a numerical model in a more expeditious and efficient way and without any computational effort.

Axisymmetric groundwater models are used for simulating radially symmetric conditions. Groundwater simulators built specifically to model axisymmetric conditions are most commonly used for simulating aquifer tests. Although some numerical models capable of simulating flow and solute transport that are developed in the cartesian coordinate system framework offer flexibility to simulate axisymmetric conditions, most of the numerical groundwater models, such as the MODFLOW family of codes, are based on structured grids in which axisymmetric flows cannot be directly simulated. Researchers in the past have provided methods to manipulate aquifer properties to mimic axisymmetric conditions. This study presents a methodology that takes advantage of the unstructured grids of MODFLOW-USG to simulate axisymmetric models within the MODFLOW framework. To develop axisymmetric models, the intercell interface area arrays of MODFLOW-USG were calculated to accurately represent coaxial cylindrical model cells. Three examples are presented to demonstrate the application of MODFLOW-USG for axisymmetric modeling: a pumping well with delayed yield effects, a vadose zone flow model simulating an infiltration basin, and a density-dependent saltwater intrusion problem for a circular island. Results were verified against analytical solutions and published numerical codes.

Sea level rise can have dramatic effects on coastal aquifers. Numerical models offer one way to study and predict these effects. A common approach is to develop and calibrate coastal aquifer models to current conditions, and then use these models to predict conditions under a variety of alternative sea level rise scenarios. Results from these scenarios can then be used to evaluate changes in groundwater flow patterns and salinity distributions in response to rising sea levels. A variety of numerical modeling programs can be used to perform these types of simulations. This presentation describes several new capabilities that are being developed for the MODFLOW suite of programs. Coastal aquifers can now be discretized using flexible meshes so that grid resolution can be focused in areas where the impacts of sea level rise are the largest. For example, the accuracy of groundwater salinization predictions can be substantially improved by increasing resolution around and beneath groundwater extraction wells. A new strategy based on the use of hydraulic head as the dependent variable, instead of freshwater head, is combined with the Newton-Raphson formulation for groundwater flow to efficiently simulate coastal aquifer water table fluctuations in response to changing sea level. These new advancements and the capability to script development and analysis of sophisticated variable-density flow and transport models with the Python programming language offer a powerful new environment for studying and predicting the effects of sea level rise on coastal aquifers.

Density-dependent flow and transport solutions for coastal saltwater intrusion investigations, analyses of fluid injection into deep brines, and studies of convective fingering and instabilities of denser fluids moving through less dense fluids typically formulate the groundwater flow equation in terms of pressure or equivalent freshwater head. A formulation of the flow equation in terms of hydraulic head is presented here as an alternative. The hydraulic-head formulation can facilitate adaptation of existing constant-density groundwater flow codes to include density-driven flow by avoiding the need to convert between freshwater head and hydraulic head within the code and by incorporating density-dependent terms as a compartmentalized "correction" to constant-density calculations already performed by the code. The hydraulic-head formulation also accommodates complexities such as unconfined groundwater flow and Newton-Raphson solution schemes more readily than the freshwater-head formulation. Simulation results are presented for four example problems solved using an implementation of the hydraulic-head formulation in MODFLOW.

A groundwater flow model is typically used to provide the flow field for conducting groundwater solute transport simulations. The advection term of the mass conserved formulation for groundwater transport assumes that the flow field is perfectly balanced and that all water flowing into a numerical grid cell is exactly balanced by outflows after accounting for sources/sinks or internal storage. However, in many complicated regional or site-scale models, there may be localized flow balance errors that may be difficult to eliminate through tighter flow convergence tolerances due to simulation time constraints or numerical limits on convergence tolerances. Thus, if water is erroneously gained or lost within a grid cell during the flow computation, the solutes within it will also be numerically affected in the associated transport simulation. Transport solutions neglect this error in groundwater flow as the transport equations that are solved assume no error in flow. This flow imbalance error can however have consequences on the transport solution ranging from unnoticeable errors in the resulting concentrations to spurious oscillations that can grow in time and hinder further solution. An approach has been suggested here, to explicitly handle these flow imbalances during mass conserved advective transport computations and report them in the corresponding transport mass balance output, as corrections that are needed to handle errors originating in the flow solution. Example problems are provided to explain the concepts and demonstrate the impacts.

The success of translating a conceptual site model for a karst site into a numerical groundwater model will depend on both the experience of the user and the capabilities and limitations of the selected computer program. Despite its numerous advantages, even MODFLOW - probably the most widely used, tested and verified modelling program currently available - has conceptual limitations that many karst hydrogeologists have to deal with on a routine basis while searching for an equivalent porous medium approach that may work. This includes assigning very high values of hydraulic conductivity to those model cells known, or suspected, to contain highly transmissive conduits, or assigning an unreasonable, very low effective porosity to the model cells with virtual 'conduits' to simulate high groundwater velocities. A new version of MODFLOW called MODFLOW-USG (UnStructured Grid) has been developed and released to the public domain. This new version retains full compatibility with previous versions of MODFLOW while taking advantage of unstructured grids and finite volume numerical solutions. It enables hydrogeologists to accurately translate even the most complex conceptual site models in karst into a numerical environment, thus eliminating the need for various surrogate modelling solutions based on an equivalent porous medium approach.

MODFLOW is a popular open-source groundwater flow model distributed by the U.S. Geological Survey. For over 30 years, the MODFLOW program has been widely used by academics, private consultants, and government scientists to accurately, reliably, and efficiently simulate groundwater flow. With time, growing interest in surface and groundwater interactions, local refinement with nested and unstructured grids, karst groundwater flow, solute transport, and saltwater intrusion, has led to the development of numerous MODFLOW versions. Although these MODFLOW versions are often based on the core MODFLOW version (presently MODFLOW-2005), there are often incompatibilities that restrict their use with other MODFLOW versions. In many cases, development of these alternative MODFLOW versions has been challenging due to the underlying program structure, which was designed for the simulation of a single groundwater flow model using a regular MODFLOW grid consisting of layers, rows, and columns. A new object-oriented program and underlying framework called MODFLOW was developed to provide a platform for supporting multiple models and multiple types of models within the same simulation. This version of MODFLOW is labeled with a 6 because it is the sixth core version of MODFLOW to be released by the USGS. In the new design, any number of models can be included in a simulation. These models can be independent of one another with no interaction, they can exchange information with one another, or they can be tightly coupled at the matrix level by adding them to the same numerical solution. Transfer of information between models is isolated to exchange objects, which allow models to be developed and used independently of one another. Within this new framework, a regional-scale groundwater model may be coupled with multiple local-scale groundwater models. Or, a surface-water flow model could be coupled to multiple groundwater flow models. The framework naturally allows for future extensions to include the simulation of solute transport. The first release of MODFLOW contains one type of hydrologic model, the Groundwater Flow (GWF) Model. The GWF Model for MODFLOW is based on a generalized control-volume finite-difference (CVFD) approach in which a cell can be hydraulically connected to any number of surrounding cells. Users can define the model grid using: 1. a regular MODFLOW grid consisting of layers, rows, and columns, 2. a layered grid defined by (x, y) vertex pairs, or 3. a general unstructured grid based on concepts developed for MODFLOW-USG. For complex problems involving water-table conditions, an optional Newton-Raphson formulation, based on the formulations in MODFLOW-NWT and MODFLOW-USG, can be activated. The GWF Model is divided into as was done in previous MODFLOW versions. A package is the part of the model that deals with a single aspect of simulation. Packages included with the GWF Model include those related to internal calculations of groundwater flow (discretization, initial conditions, hydraulic conductance, and storage), stress packages (constant heads, wells, recharge, rivers, general head boundaries, drains, and evapotranspiration), and advanced stress packages (streamflow routing, lakes, multi-aquifer wells, and unsaturated zone flow). An additional package is also available for moving water available in one package into the individual features of the advanced stress packages. The GWF Model also has packages for obtaining and controlling output from the model. A new groundwater flow formulation was developed specifically for the GWF Model in MODFLOW 6. This new flow formulation is called the XT3D option. The XT3D option extends the capabilities of MODFLOW by enabling simulation of fully three-dimensional anisotropy on regular or irregular grids in a way that properly takes into account the full, three-dimensional hydraulic conductivity tensor. It can also improve the accuracy of groundwater flow simulations in cases in which the model grid violates certain geometric requirements. Thus, the XT3D option is an alternative to the Ghost-Node Correction (GNC) Package, which was developed for MODFLOW-USG. In addition to the many new or redesigned capabilities, the MODFLOW input structure has also been redesigned. Within package input files, information is divided into blocks and informative keywords are used to label numeric data and activate options. This new input structure was designed to make it easier for users to adjust simulation options in an intuitive manner, reduce user input errors, and allow new capabilities to be added without causing problems with backward compatibility. This first MODFLOW release with the GWF Model synthesizes many of the capabilities in existing MODFLOW variants: 1. MODFLOW-2005 the GWF Model contains revisions of the commonly used flow packages, stress packages, and advanced stress packages 2. MODFLOW-NWT the GWF Model supports an optional Newton-Raphson approach for water table aquifers 3. MODFLOW-USG GWF Models can be developed using regular MODFLOW grids or unstructured grids 4. MODFLOW-LGR any number of GWF Models can be specified for a single simulation; these GWF Models are tightly coupled at the matrix level 5. MODFLOW-CDSS multiple stress and advanced stress packages of the same type can be specified for a single GWF Model For MODFLOW users with existing models, the MODFLOW distribution includes a conversion program that will translate a MODFLOW-2005, MODFLOW-NWT, or MODFLOW-LGR2 model into the MODFLOW format.

First posted August 10, 2017 For additional information, contact: Office of Groundwater U.S. Geological Survey Mail Stop 411 12201 Sunrise Valley Drive Reston, VA 20192 This report documents the Groundwater Flow (GWF) Model for a new version of MODFLOW called MODFLOW 6. The GWF Model for MODFLOW 6 is based on a generalized control-volume finite-difference approach in which a cell can be hydraulically connected to any number of surrounding cells. Users can define the model grid using one of three discretization packages, including (1) a structured discretization package for defining regular MODFLOW grids consisting of layers, rows, and columns, (2) a discretization by ver­tices package for defining layered unstructured grids consisting of layers and cells, and (3) a general unstruc­tured discretization package for defining flexible grids comprised of cells and their connection properties. For layered grids, a new capability is available for removing thin cells and vertically connecting cells overlying and underlying the thin cells. For complex problems involving water-table conditions, an optional Newton-Raphson formulation, based on the formulations in MODFLOW-NWT and MODFLOW-USG, can be acti­vated. Use of the Newton-Raphson formulation will often improve model convergence and allow solutions to be obtained for difficult problems that cannot be solved using the traditional wetting and drying approach. The GWF Model is divided into “packages,” as was done in previous MODFLOW versions. A package is the part of the model that deals with a single aspect of simulation. Packages included with the GWF Model include those related to internal calculations of groundwater flow (discretization, initial conditions, hydraulic conduc­tance, and storage), stress packages (constant heads, wells, recharge, rivers, general head boundaries, drains, and evapotranspiration), and advanced stress packages (streamflow routing, lakes, multi-aquifer wells, and unsaturated zone flow). An additional package is also available for moving water available in one package into the individual features of the advanced stress packages. The GWF Model also has packages for obtaining and controlling output from the model. This report includes detailed explanations of physical and mathematical concepts on which the GWF Model and its packages are based.Like its predecessors, MODFLOW 6 is based on a highly modular structure; however, this structure has been extended into an object-oriented framework. The framework includes a robust and generalized numeri­cal solution object, which can be used to solve many different types of models. The numerical solution object has several different matrix preconditioning options as well as several methods for solving the linear system of equations. In this new framework, the GWF Model itself is an object as are each of the GWF Model packages. A benefit of the object-oriented structure is that multiple objects of the same type can be used in a single sim­ulation. Thus, a single forward run with MODFLOW 6 may contain multiple GWF Models. GWF Models can be hydraulically connected using GWF-GWF Exchange objects. Connecting GWF models in different ways permits the user to utilize a local grid refinement strategy consisting of parent and child models or to couple adjacent GWF Models. An advantage of the approach implemented in MODFLOW 6 is that multiple models and their exchanges can be incorporated into a single numerical solution object. With this design, models can be tightly coupled at the matrix level.

Production of Coalbed Methane (CBM) requires extraction of large quantities of groundwater. To date, standard groundwater flow simulators have mostly been used to assess the impact of this extraction on regional groundwater systems. Recent research has demonstrated that predictions of regional impact assessment made by such models may be seriously compromised unless account is taken of the presence of a gas phase near extraction wells. At the same time, CBM impact assessment must accommodate the traditional requirements of regional groundwater modelling. These include representation of surficial groundwater processes and up-scaled rock properties as well as the need for calibration and predictive uncertainty quantification. The study documented herein (1) quantifies errors in regional drawdown predictions incurred through neglect of the presence of a gas phase near CBM extraction centres, and (2) evaluates the extent to which these errors can be mitigated by simulating near-well desaturation using a modified Richards equation formulation within a standard groundwater flow simulator. Two synthetic examples are provided to quantify the impact of the gas phase and verify the proposed modelling approach (implemented in MODFLOW-USG) against rigorous multiphase flow simulations (undertaken using ECLIPSE*). ECLIPSE simulations demonstrate convergence towards a time-asymptotic relationship between water saturation and pressure. This relationship can be approximated using a slightly modified van Genuchten function. Where this function is employed in combination with the modified Richards equation strategy to accommodate near-well desaturation, errors in predicted drawdown are reduced significantly, including in cases where complexities such as sloping coal layers are introduced to the model domain (the latter promoting buoyancy-driven movement of gas). Sensitivity analyses further indicate that only the general properties of the employed desaturation function need to be respected to significantly reduce errors in regional drawdown predictions that would arise if the presence of the near-well gas phase was ignored. These properties can be inferred from reservoir properties and from the outcomes of reservoir model simulations that are available at local CBM operation sites. Crown Copyright (C) 2015 Published by Elsevier B.V. All rights reserved.

Infiltration along ephemeral channels plays an important role in groundwater recharge in arid regions. A model is presented for estimating spatial variability of seepage due to streambed heterogeneity along channels based on measurements of streamflow-front velocities in initially dry channels. The diffusion-wave approximation to the Saint-Venant equations, coupled with Philip's equation for infiltration, is connected to the groundwater model MODFLOW and is calibrated by adjusting the saturated hydraulic conductivity of the channel bed. The model is applied to portions of two large water delivery canals, which serve as proxies for natural ephemeral streams. Estimated seepage rates compare well with previously published values. Possible sources of error stem from uncertainty in Manning's roughness coefficients, soil hydraulic properties and channel geometry. Model performance would be most improved through more frequent longitudinal estimates of channel geometry and thalweg elevation, and with measurements of stream stage over time to constrain wave timing and shape. This model is a potentially valuable tool for estimating spatial variability in longitudinal seepage along intermittent and ephemeral channels over a wide range of bed slopes and the influence of seepage rates on groundwater levels. Copyright (c) 2012 John Wiley & Sons, Ltd.