Recharge estimation for transient ground water modeling

Reliable ground water models require both an accurate physical representation of the system and appropriate boundary conditions. While physical attributes are generally considered static, boundary conditions, such as ground water recharge rates, can be highly variable in both space and time. A practical methodology incorporating the hydrologic model HELP3 in conjunction with a geographic information system was developed to generate a physically based and highly detailed recharge boundary condition for ground water modeling. The approach uses daily precipitation and temperature records in addition to land use/land cover and soils data. The importance of the method in transient ground water modeling is demonstrated by applying it to a MODFLOW modeling study in New Jersey. In addition to improved model calibration, the results from the study clearly indicate the importance of using a physically based and highly detailed recharge boundary condition in ground water quality modeling, where the detailed knowledge of the evolution of the ground water flowpaths is imperative. The simulated water table is within 0.5 m of the observed values using the method, while the water levels can differ by as much as 2 m using uniform recharge conditions. The results also show that the combination of temperature and precipitation plays an important role in the amount and timing of recharge in cooler climates. A sensitivity analysis further reveals that increasing the leaf area index, the evaporative zone depth, or the curve number in the model will result in decreased recharge rates over time, with the curve number having the greatest impact.     

Determining extreme parameter correlation in ground water models

In ground water flow system models with hydraulic-head observations but without significant imposed or observed flows, extreme parameter correlation generally exists. As a result, hydraulic conductivity and recharge parameters cannot be uniquely estimated. In complicated problems, such correlation can go undetected even by experienced modelers. Extreme parameter correlation can be detected using parameter correlation coefficients, but their utility depends on the presence of sufficient, but not excessive, numerical imprecision of the sensitivities, such as round-off error. This work investigates the information that can be obtained from parameter correlation coefficients in the presence of different levels of numerical imprecision, and compares it to the information provided by an alternative method called the singular value decomposition (SVD). Results suggest that (1) calculated correlation coefficients with absolute values that round to 1.00 were good indicators of extreme parameter correlation, but smaller values were not necessarily good indicators of lack of correlation and resulting unique parameter estimates; (2) the SVD may be more difficult to interpret than parameter correlation coefficients, but it required sensitivities that were one to two significant digits less accurate than those that required using parameter correlation coefficients; and (3) both the SVD and parameter correlation coefficients identified extremely correlated parameters better when the parameters were more equally sensitive. When the statistical measures fail, parameter correlation can be identified only by the tedious process of executing regression using different sets of starting values, or, in some circumstances, through graphs of the objective function.     

Adaptive underrelaxation of Picard iterations in ground water models

This methods note examines the use of adaptive underrelaxation of Picard iterations to accelerate the solution convergence for nonlinear ground water flow problems. Ground water problems are nonlinear when drains, phreatophytes, stream aquifer, and similar features are simulated. Typically, simple Picard iterations are used to address such nonlinear problems. Nevertheless, the convergence rate can be slow, or convergence cannot be obtained. However, convergence often can be accelerated using Picard iterations with adaptive underrelaxation, and convergence often can be obtained where it otherwise would not occur.     

Regional estimation of total recharge to ground water in Nebraska

Naturally occurring long-term mean annual recharge to ground water in Nebraska was estimated by a novel water-balance approach. This approach uses geographic information systems (GIS) layers of land cover, elevation of land and ground water surfaces, base recharge, and the recharge potential in combination with monthly climatic data. Long-term mean recharge > 140 mm per year was estimated in eastern Nebraska, having the highest annual precipitation rates within the state, along the Elkhorn, Platte, Missouri, and Big Nemaha River valleys where ground water is very close to the surface. Similarly high recharge values were obtained for the Sand Hills sections of the North and Middle Loup, as well as Cedar River and Beaver Creek valleys due to high infiltration rates of the sandy soil in the area. The westernmost and southwesternmost parts of the state were estimated to typically receive < 30 mm of recharge a year.     

The practical use of simplicity in developing ground water models

The advantages of starting with simple models and building complexity slowly can be significant in the development of ground water models. In many circumstances, simpler models are characterized by fewer defined parameters and shorter execution times. In this work, the number of parameters is used as the primary measure of simplicity and complexity; the advantages of shorter execution times also are considered. The ideas are presented in the context of constructing ground water models but are applicable to many fields. Simplicity first is put in perspective as part of the entire modeling process using 14 guidelines for effective model calibration. It is noted that neither very simple nor very complex models generally produce the most accurate predictions and that determining the appropriate level of complexity is an ill-defined process. It is suggested that a thorough evaluation of observation errors is essential to model development. Finally, specific ways are discussed to design useful ground water models that have fewer parameters and shorter execution times.     

A methodology for validating numerical ground water models

Ground water validation is one of the most challenging issues facing modelers and hydrogeologists. Increased complexity in ground water models has created a gap between model predictions and the ability to validate or build confidence in predictions. Specific procedures and tests that can be easily adapted and applied to determine the validity of site-specific ground water models do not exist. This is true for both deterministic and stochastic models, with stochastic models posing the more difficult validation problem. The objective of this paper is to propose a general validation approach that addresses important issues recognized in previous validation studies, conferences, and symposia. The proposed method links the processes for building, calibrating, evaluating, and validating models in an iterative loop. The approach focuses on using collected validation data to reduce uncertainty in the model and narrow the range of possible outcomes. This method is designed for stochastic numerical models utilizing Monte Carlo simulation approaches, but it can be easily adapted for deterministic models. The proposed methodology relies on the premise that absolute validity is not theoretically possible, nor is it a regulatory requirement. Rather, the proposed methodology highlights the importance of testing various aspects of the model and using diverse statistical tools for rigorous checking and confidence building in the model and its predictions. It is this confidence that will encourage regulators and the public to accept decisions based on the model predictions. This validation approach will be applied to a model, described in this paper, dealing with an underground nuclear test site in rural Nevada.     

Testing alternative ground water models using cross-validation and other methods

Many methods can be used to test alternative ground water models. Of concern in this work are methods able to (1) rank alternative models (also called model discrimination) and (2) identify observations important to parameter estimates and predictions (equivalent to the purpose served by some types of sensitivity analysis). Some of the measures investigated are computationally efficient; others are computationally demanding. The latter are generally needed to account for model nonlinearity. The efficient model discrimination methods investigated include the information criteria: the corrected Akaike information criterion, Bayesian information criterion, and generalized cross-validation. The efficient sensitivity analysis measures used are dimensionless scaled sensitivity (DSS), composite scaled sensitivity, and parameter correlation coefficient (PCC); the other statistics are DFBETAS, Cook's D, and observation-prediction statistic. Acronyms are explained in the introduction. Cross-validation (CV) is a computationally intensive nonlinear method that is used for both model discrimination and sensitivity analysis. The methods are tested using up to five alternative parsimoniously constructed models of the ground water system of the Maggia Valley in southern Switzerland. The alternative models differ in their representation of hydraulic conductivity. A new method for graphically representing CV and sensitivity analysis results for complex models is presented and used to evaluate the utility of the efficient statistics. The results indicate that for model selection, the information criteria produce similar results at much smaller computational cost than CV. For identifying important observations, the only obviously inferior linear measure is DSS; the poor performance was expected because DSS does not include the effects of parameter correlation and PCC reveals large parameter correlations.     

PRO-GRADE: GIS toolkits for ground water recharge and discharge estimation

PRO-GRADE is an ESRI ArcGIS 9.2 plug-in package that consists of two separate toolkits: (1) the p attern r ecognition o rganizer for g eographic i nformation s ystem (PRO-GIS) and (2) the g round water r echarge a nd d ischarge e stimator for GIS (GRADE-GIS). PRO-GIS is a collection of several existing image-processing algorithms into one user interface to offer the flexibility to extract spatial patterns according to the user's needs. GRADE-GIS is a ground water recharge and discharge estimation interface using a mass balance method that requires only hydraulic conductivity, water table, and bedrock elevation data for simulating two-dimensional steady-state unconfined aquifers. PRO-GRADE was developed to assist ongoing assessments of the water resources in Illinois and Wisconsin, and is being used to assist several ground water resource studies in several locations in the United States. The advantage of using PRO-GRADE is to enable fast production of initial recharge and discharge maps that can be further enhanced by using a follow-up ground water flow model with parameter estimation codes. PRO-GRADE leverages ArcGIS to provide a computer-assisted framework to support expert judgment in order to efficiently select alternative recharge and discharge maps that can be used as (1) guidelines for field study planning and decision making; (2) initial conditions for numerical simulation; and (3) screening for alternative model selection and prediction/parameter uncertainty evaluation. In addition, PRO-GRADE allows for more easy and rapid correlation of those maps with other hydrologically relevant geospatial data.     

Regional estimation of base recharge to ground water using water balance and a base-flow index

Naturally occurring long-term mean annual base recharge to ground water in Nebraska was estimated with the help of a water-balance approach and an objective automated technique for base-flow separation involving minimal parameter-optimization requirements. Base recharge is equal to total recharge minus the amount of evapotranspiration coming directly from ground water. The estimation of evapotranspiration in the water-balance equation avoids the need to specify a contributing drainage area for ground water, which in certain cases may be considerably different from the drainage area for surface runoff. Evapotranspiration was calculated by the WREVAP model at the Solar and Meteorological Surface Observation Network (SAMSON) sites. Long-term mean annual base recharge was derived by determining the product of estimated long-term mean annual runoff (the difference between precipitation and evapotranspiration) and the base-flow index (BFI). The BFI was calculated from discharge data obtained from the U.S. Geological Survey's gauging stations in Nebraska. Mapping was achieved by using geographic information systems (GIS) and geostatistics. This approach is best suited for regional-scale applications. It does not require complex hydrogeologic modeling nor detailed knowledge of soil characteristics, vegetation cover, or land-use practices. Long-term mean annual base recharge rates in excess of 110 mm/year resulted in the extreme eastern part of Nebraska. The western portion of the state expressed rates of only 15 to 20 mm annually, while the Sandhills region of north-central Nebraska was estimated to receive twice as much base recharge (40 to 50 mm/year) as areas south of it.     

Validation of numerical ground water models used to guide decision making

Many sites of ground water contamination rely heavily on complex numerical models of flow and transport to develop closure plans. This complexity has created a need for tools and approaches that can build confidence in model predictions and provide evidence that these predictions are sufficient for decision making. Confidence building is a long-term, iterative process and the author believes that this process should be termed model validation. Model validation is a process, not an end result. That is, the process of model validation cannot ensure acceptable prediction or quality of the model. Rather, it provides an important safeguard against faulty models or inadequately developed and tested models. If model results become the basis for decision making, then the validation process provides evidence that the model is valid for making decisions (not necessarily a true representation of reality). Validation, verification, and confirmation are concepts associated with ground water numerical models that not only do not represent established and generally accepted practices, but there is not even widespread agreement on the meaning of the terms as applied to models. This paper presents a review of model validation studies that pertain to ground water flow and transport modeling. Definitions, literature debates, previously proposed validation strategies, and conferences and symposia that focused on subsurface model validation are reviewed and discussed. The review is general and focuses on site-specific, predictive ground water models used for making decisions regarding remediation activities and site closure. The aim is to provide a reasonable starting point for hydrogeologists facing model validation for ground water systems, thus saving a significant amount of time, effort, and cost. This review is also aimed at reviving the issue of model validation in the hydrogeologic community and stimulating the thinking of researchers and practitioners to develop practical and efficient tools for evaluating and refining ground water predictive models.     

Estimating ground water recharge using flow models of perched karstic aquifers

The fraction of rain that is annually recharged to ground water is a function of the transient quantities of precipitation (wet vs. dry years) as well as other meteorological and geologic factors, and thus it is very difficult to estimate. In this study, we have used long records (20 to 30 years) of precipitation and spring discharge to reconstruct the transient character of yearly recharge. These data sets were used to calibrate numerical ground water flow models on the less than 3 km(2) scale for four separate perched karstic aquifers in the Judean and Samarian Mountains of Israel. The stratification and karstic character of the local carbonate rock aquifers cause ground water to flow through discrete dissolution channels and to discharge at isolated springs. An innovative, dual-porosity approach was used where a finite-difference solution simulates flow in the rock matrix, while the karstic channels are simulated using computationally simple drains. Perched conditions are also simulated innovatively using MODFLOW by treating the bottom unsaturated layer as if it is saturated, but by assuming zero pressure head throughout the "unsaturated" layer. Best fitting between measured and computed spring hydrograph data has allowed us to develop a set of empirical functions relating measured precipitation to recharge to the aquifer. The generic methodology presented gives insight into the suspected changes in aquifer recharge rates between particularly wet or dry years.     

An assessment of aquifer storage recovery using ground water flow models

Owing to increased demands on ground water accompanied by increased drawdowns, technologies that use recharge options, such as aquifer storage recovery (ASR), are being used to optimize available water resources and reduce adverse effects of pumping. In this paper, three representative ground water flow models were created to assess the impact of hydrogeologic and operational parameters/factors on recovery efficiency of ASR systems. Flow/particle tracking and solute transport models were used to track the movement of water during injection, storage, and recovery. Results from particle tracking models consistently produced higher recovery efficiency than the solute transport models for the parameters/properties examined because the particle tracking models neglected mixing of the injected and ambient water. Mixing between injected and ambient water affected recovery efficiency. Results from this study demonstrate the interactions between hydrogeologic and operational parameters on predictions of recovery efficiency. These interactions are best simulated using coupled numerical ground water flow and transport models that include the effects of mixing of injected water and ambient ground water.     

A digital procedure for ground water recharge and discharge pattern recognition and rate estimation

A digital procedure to estimate recharge/discharge rates that requires relatively short preparation time and uses readily available data was applied to a setting in central Wisconsin. The method requires only measurements of the water table, fluxes such as stream baseflows, bottom of the system, and hydraulic conductivity to delineate approximate recharge/discharge zones and to estimate rates. The method uses interpolation of the water table surface, recharge/discharge mapping, pattern recognition, and a parameter estimation model. The surface interpolator used is based on the theory of radial basis functions with thin-plate splines. The recharge/discharge mapping is based on a mass-balance calculation performed using MODFLOW. The results of the recharge/discharge mapping are critically dependent on the accuracy of the water table interpolation and the accuracy and number of water table measurements. The recharge pattern recognition is performed with the help of a graphical user interface (GUI) program based on several algorithms used in image processing. Pattern recognition is needed to identify the recharge/discharge zonations and zone the results of the mapping method. The parameter estimation program UCODE calculates the parameter values that provide a best fit between simulated heads and flows and calibration head-and-flow targets. A model of the Buena Vista Ground Water Basin in the Central Sand Plains of Wisconsin is used to demonstrate the procedure.     

Neural estimation of strong ground motion duration

This paper presents and discusses the use of neural networks to determine strong ground motion duration. Accelerometric data recorded in the Mexican cities of Puebla and Oaxaca are used to develop a neural model that predicts this duration in terms of the magnitude, epicenter distance, focal depth, soil characterization and azimuth. According to the above the neural model considers the effect of the seismogenic zone and the contribution of soil type to the duration of strong ground motion. The final scheme permits a direct estimation of the duration since it requires easy-to-obtain variables and does not have restrictive hypothesis. The results presented in this paper indicate that the soft computing alternative, via the neural model, is a reliable recording-based approach to explore and to quantify the effect of seismic and site conditions on duration estimation. An essential and significant aspect of this new model is that, while being extremely simple, it also provides estimates of strong ground motions duration with remarkable accuracy. Additional but important side benefits arising from the model’s simplicity are the natural separation of source, path, and site effects and the accompanying computational efficiency.     

Site-dependent models of earthquake ground motion

Random vibration analyses of structural systems subjected to seismic loading are dependent upon the characterization of earthquake ground motion as a stochastic process. The response of structural systems to earthquakes is dependent strongly on the local geological conditions, which should be incorporated into seismological models of ground motion. In the study presented herein, three previously developed ground-motion models are adapted to incorporate site-dependent characteristics. Records obtained from two recording stations in California are used as a basis for the ground-motion models. Single-degree-of-freedom (SDOF) oscillators are subjected to ensembles of accelerograms generated from these models, and both elastic and inelastic response are considered. Response statistics are compared to those generated by the analysis of structural response to ensembles of recorded motion from the two sites. The important features of the ground motion for effective reproduction of response statistics are identified, and observations are made on the sensitivity of specific response parameters to site-dependent characteristics of the ground motion.     

Uncertainty estimates for surface runoff models

A lower bound for variance in surface runoff modelling estimates is advanced. The bound is derived using a linear unit hydrograph approach which utilizes a discretization of the catchment into an arbitrary number of subareas, a linear routing technique for channel flow effects, a variable effective rainfall distribution over the catchment, and calibration parameter distributions developed in correlating rainfall-runoff data by the model. The uncertainty bound reflects the dominating influence of the unknown rainfall distribution over the catchment and is expressed as a distribution function that can be reduced only by supplying additional rainfall-runoff data. It is recommended that this uncertainty distribution in modelling results be included in flood control design studies in order to incorporate a prescribed level of confidence in flood protection facilities.     

A particle-tracking scheme for simulating pathlines in coupled surface-subsurface flows

A Lagrangian particle tracking scheme has been extended to simulate advective transport through coupled surface and subsurface flows. This extended scheme assumes a two-dimensional flow field for the overland domain and a three-dimensional flow field for the subsurface domain. Moreover it is assumed that the flow fields are simulated by a cell centered finite difference method. Pathlines in both the subsurface and the overland domain are simulated by classical particle tracking methods. Exchange of particles between the two domains is simulated by newly-developed algorithms presented in this study. Different algorithms are used depending on the direction of the exchange across the interface between the two domains. In the subsurface domain knowledge about a particle’s pathline is enough to detect a transfer to the surface domain and the solution is straightforward. However, in the two-dimensional overland domain pathlines are parallel to the land surface. Therefore the velocity field in the overland domain alone cannot be used to detect a transfer to the subsurface. We propose a relatively simple algorithm to estimate the probability of transfer to the subsurface domain. It is shown that this algorithm can also be used to handle the transfer from the overland domain to the atmosphere domain. The algorithm to estimate the transfer probabilities is based on the mass balance of water on a streamtube aligned with the particle’s pathline. This newly developed technique ensures that transit time distributions can be simulated accurately. These new relationships are implemented in an existing particle tracking code and are verified using analytical solutions for transit times.