Groundwater Model Calibration Research Articles

A comprehensive framework for stochastic calibration and sensitivity analysis of large-scale groundwater models

Abstract. We introduce a comprehensive and robust theoretical framework and operational workflow that can be employed to enhance our understanding, modeling and management capability of complex heterogeneous large-scale groundwater systems. Our framework encapsulates key components such as the three-dimensional nature of groundwater flows, river–aquifer interactions, probabilistic reconstruction of three-dimensional spatial distributions of geomaterials and associated properties across the subsurface, multi-objective optimization for model parameter estimation through stochastic calibration, and informed global sensitivity analysis (GSA). By integrating these components, we effectively consider the inherent uncertainty associated with subsurface system characterizations as well as their interactions with surface waterbodies. The approach enables us to identify parameters impacting diverse system responses. By employing a coevolutionary optimization algorithm, we ensure efficient model parameterization, facilitating simultaneous and informed optimization of the defined objective functions. Additionally, estimation of parameter uncertainty naturally leads to quantification of uncertainty in system responses. The methodology is designed to increase our knowledge of the dynamics of large-scale groundwater systems. It also has the potential to guide future data acquisition campaigns through an informed global sensitivity analysis. We demonstrate the effectiveness of our proposed methodology by applying it to the largest groundwater system in Italy. We address the challenges posed by the characterization of the heterogeneous spatial distribution of subsurface attributes across large-scale three-dimensional domains upon incorporating a recent probabilistic hydrogeological reconstruction specific to the study case. The system considered faces multiple challenges, including groundwater contamination, seawater intrusion, and water scarcity. Our study offers a promising modeling strategy applicable to large-scale subsurface systems and valuable insights into groundwater flow patterns that can then inform effective system management.

Numerical modeling of groundwater flow and nitrate transport in karst and wastewater irrigated agricultural and forest landscapes, PA, USA

AbstractUnderstanding the influence of treated wastewater reuse for irrigation on nitrate groundwater contamination in karstic landscape is critical for developing long‐term management plans and assessing the sustainability of these practices. The Pennsylvania State University has been operating a secondary treated wastewater effluent irrigation site—the “Living Filter”—since the 1960s. The objective of this study is to develop a regional groundwater flow and nitrate transport model to assess the impact of treated wastewater and disposal of large volumes of wastewater for irrigation of agricultural and forested lands on nitrate groundwater contamination in a karst landscape in the Spring Creek Watershed in central Pennsylvania. The modular three‐dimensional finite‐difference ground‐water flow model (MODFLOW) was used to construct the groundwater flow model of the local aquifer and simulate three‐dimensional regional changes in the water table over time. The simulation of nitrate fate and transport in the aquifer was conducted by developing a groundwater mass transport model using MODFLOW and a modular three‐dimensional transport multi‐species model (MT3DMS). The calibrated groundwater model yielded a root mean square error of 5.41 ft for the area of irrigation and 24.67 ft for the entire modeled area. Results of the calibrated model show the extension of a nitrate plume over most of the study area with concentrations below 10 mg·L−1 under the current 2015 setting and the long‐term 2035 scenario of increased irrigation rates. The nitrate migration was dominated by the presence of two high conductivity groundwater troughs adjacent to the study area resulting in a near steady‐state total mass since the 1990s.

Calibration of groundwater seepage against the spatial distribution of the stream network to assess catchment-scale hydraulic properties

Abstract. The assessment of effective hydraulic properties at the catchment scale, i.e., hydraulic conductivity (K) and transmissivity (T), is particularly challenging due to the sparse availability of hydrological monitoring systems through stream gauges and boreholes. To overcome this challenge, we propose a calibration methodology which only considers information from a digital elevation model (DEM) and the spatial distribution of the stream network. The methodology is built on the assumption that the groundwater system is the main driver controlling the stream density and extension, where the perennial stream network reflects the intersection of the groundwater table with the topography. Indeed, the groundwater seepage at the surface is primarily controlled by the topography, the aquifer thickness and the dimensionless parameter K/R, where R is the average recharge rate. Here, we use a process-based and parsimonious 3D groundwater flow model to calibrate K/R by minimizing the relative distances between the observed and the simulated stream network generated from groundwater seepage zones. By deploying the methodology in 24 selected headwater catchments located in northwestern France, we demonstrate that the method successfully predicts the stream network extent for 80 % of the cases. Results show a high sensitivity of K/R to the extension of the low-order streams and limited impacts of the DEM resolution as long the DEM remains consistent with the stream network observations. By assuming an average recharge rate, we found that effective K values vary between 1.0×10-5 and 1.1×10-4 m s−1, in agreement with local estimates derived from hydraulic tests and independent calibrated groundwater model. With the emergence of global remote-sensing databases compiling information on high-resolution DEM and stream networks, this approach provides new opportunities to assess hydraulic properties of unconfined aquifers in ungauged basins.

Suitable sites for groundwater development: a capture map-based approach integrated with weighted overlay analysis

Abstract A novel methodology for suitable site selection for groundwater development based on river capture, pumping cost, and groundwater potential has been proposed for better groundwater utilisation. River capture and cost map have been generated from a calibrated groundwater model, simulated with forecasted hydrological time series data. The groundwater potential has been calculated with weighted overlay analysis. These three variables have been used to classify the model domain into five zones of groundwater development by K-Means clustering. The area with lower river capture, low cost of pumping, and high groundwater potential is found to be the best location for groundwater extraction. The methodology has been applied to the lower Ain River basin, in France.

Groundwater modeling of the Silala basin and impacts of channelization

AbstractA key issue in the legal case before the International Court of Justice between Bolivia and Chile concerned artificial channels in the headwater wetlands of the Silala River in Bolivia, and their impact on the river and groundwater flow rates across the international border into Chile. In pleadings before the Court, Bolivia had claimed up to 40% changes in river flow, whereas Chile had maintained that these were much smaller. This paper reviews construction of a numerical groundwater model to quantify this impact, based on the hydrogeological conceptual model presented earlier in this volume. The groundwater model domain covers the Silala River topographical catchment, plus three adjacent closed (endorheic) topographical catchments to the northeast, which together form the groundwater catchment of the Silala River. The geometry was based on a 3D geological model, constructed based on knowledge gained from several geological mapping investigations. Surface water flows were simulated within a margin of error less than the range of uncertainty associated with the flow measurements. The resulting calibrated groundwater model was used to simulate scenarios corresponding to (a) the removal of channels in the Bolivian wetlands and (b) backfilling of the channels and a small increase in ground elevation (postulated by Bolivia due to long term peat accumulation). The results of these scenarios indicate that the measures would result in a small reduction to the surface water flow over the border to Chile; for Scenario A less than 1% change in surface flows, and for Scenario B less than 3%.This article is categorized under: Human Water > Rights to Water Science of Water > Hydrological Processes Human Water > Water Governance

Evaluating the contribution of satellite-derived evapotranspiration in the calibration of numerical groundwater models in remote zones using the EEFlux tool

Assessment of groundwater resources is crucial for developing water management practices for its sustainable exploitation, both for current and future needs. Numerical models are useful tools for such purpose. However, the lack of continuous monitoring networks, mainly due to difficult access to some remote locations, poses a challenge in developing and calibrating groundwater models. Remote sensing offers an alternative for acquiring information on hydrological and climatic variables at multiple spatiotemporal scales that has the potential to strengthen groundwater modeling. The aim of this study is to develop a methodology that uses remote sensing products to support model calibration. With this aim, we used the Parameter Estimation software (PEST) to calibrate a hydrogeological model of an unexploited basin located in the arid Chilean Altiplano using observed groundwater levels and evapotranspiration (ET) derived from the Earth Engine Evapotranspiration Flux (EEFlux) tool as observations. Our results show that the best model calibration is achieved using both EEFlux-ET and heads as observations to calibrate the hydraulic properties (normalized root mean square error = 4.1 %). We analyzed the effect of EEFlux-ET on the calibration of these properties and found a direct effect on specific yield parameters, which regulate the fluctuations of the water table over time. Incorporating EEFlux-ET estimates in the calibration resulted in lower values of specific yield across the aquifer. Therefore, incorporating remotely sensed ET as observations in the calibration of the groundwater model contributes to a better simulation of the spatiotemporal head variations in the basin.

A robust objective function for calibration of groundwater models in light of deficiencies of model structure and observations

• Need for robust objective functions in model calibration and parameter estimation. • Common squared error_based criteria have issues with outliers and extreme values. • Suggestion: robust objective function based on CRPS as heuristic norm. • Comparison of CRPS with other performance criteria in groundwater model calibration. Groundwater models require parameter optimization based on the minimization of objective functions describing, for example, the residual between observed and simulated groundwater head. At larger scales, constraining these models requires large datasets of groundwater head observations. These observations are typically only available from databases comprised of varying quality data from a variety of sources and will be associated with unknown observational uncertainty. At the same time the model structure, especially the hydrogeological description, will inevitably be a simplification of the complex natural system. As a result, calibration of groundwater models often results in parameter compensation for model structural deficiency, or can be affected by observation errors. This problem can be amplified by the application of common squared error-based performance criteria, which are most sensitive to the largest errors. In our context, we assume that the residuals that remain large during the optimization process likely do so because of either model structural error or observation error. Based on this assumption it is desirable to design an objective function that is less sensitive to these large residuals of low probability, and instead favours the majority of observations that can fit the given model structure and likely are free of large observation errors. We suggest a Continuous Ranked Probability Score (CRPS) based objective function that limits the influence of large residuals in the optimization process as the metric puts more emphasis on the position of the residual along the cumulative distribution function than on the magnitude of the residual. The CRPS-based objective function was applied in the calibration of regional-scale coupled surface-groundwater models and compared to conventional objective functions based on mean of squared, absolute and root errors, using synthetic as well as real observations. The optimization tests illustrated that the novel CRPS-based objective function successfully limited the dominance of large residuals in the optimization process and consistently reduced overall bias.

Informing hydrogeological models with remotely sensed evapotranspiration

Remotely sensed evapotranspiration (ET) rates can provide an additional constraint on the calibration of groundwater models beyond typically-used water table (WT) level observations. The value of this constraint, measured in terms of reductions in model error, however, is expected to vary with the method by which it is imposed and by how closely the ET flux is dependant to groundwater levels. To investigate this variability, four silvicultural sites with different access to groundwater were modeled under three different model-data configurations. A benchmark model that used only WT levels for calibration was compared to two alternatives: one in which satellite remotely sensed ET rates from MODIS-CMRSET were also included in model calibration, and one in which the satellite ET data were assimilated, through the Ensemble Kalman Filter, into the model. Large error reductions in ET flux outputs were achieved when CMRSET data were used to calibrate the model. Assimilation of CMRSET data further improved the model performance statistics where the WT was < 6.5 m deep. It is advantageous to use spatially distributed actual ET data to calibrate groundwater models where it is available. In situations where vegetation has direct access to groundwater, assimilation of ET observations is likely to improve model performance.

Net Inflow: An Important Target on the Path to Aquifer Sustainability.

Aquifers supporting irrigated agriculture are a resource of global importance. Many of these systems, however, are experiencing significant pumping-induced stress that threatens their continued viability as a water source for irrigation. Reductions in pumping are often the only option to extend the lifespans of these aquifers and the agricultural production they support. The impact of reductions depends on a quantity known as "net inflow" or "capture." We use data from a network of wells in the western Kansas portions of the High Plains aquifer in the central United States to demonstrate the importance of net inflow, how it can be estimated in the field, how it might vary in response to pumping reductions, and why use of "net inflow" may be preferred over "capture" in certain contexts. Net inflow has remained approximately constant over much of western Kansas for at least the last 15 to 25 years, thereby allowing it to serve as a target for sustainability efforts. The percent pumping reduction required to reach net inflow (i.e., stabilize water levels for the near term [years to a few decades]) can vary greatly over this region, which has important implications for groundwater management. However, the reduction does appear practically achievable (less than 30%) in many areas. The field-determined net inflow can play an important role in calibration of regional groundwater models; failure to reproduce its magnitude and temporal variations should prompt further calibration. Although net inflow is a universally applicable concept, the reliability of field estimates is greatest in seasonally pumped aquifers.

Coastal groundwater model calibration using filtered and amplified hydraulic information retained in the freshwater–saltwater interface

Coastal groundwater model calibration using filtered and amplified hydraulic information retained in the freshwater–saltwater interface

Simplifying the Groundwater Model Interface to Provide Enhanced Decision Support

AbstractGroundwater modeling can be made more accessible to stakeholders and water managers through interface tools that prepare model inputs and create graphical outputs. An application of this improved modeling technique is presented here, where a model interface tool, called the Groundwater Management Tool (GMT) was developed for the widely used MODFLOW model. The GMT facilitates the evaluation of various water use scenarios for those who have limited time for and/or experience in modifying groundwater model inputs and outputs, yet who need the complexity of a three‐dimensional numerical groundwater model to adequately represent the hydrologic system of interest. The impacts of new pumping wells and altered pumping regimes in existing wells are easily explored through the current version of this interface tool. The modeling technique employed by the GMT provides water managers with the ability to efficiently leverage the decision‐support power and confidence provided by a calibrated groundwater model without the time‐consuming processing of model input and output files, and thereby increases the value and utility provided by the model in supporting water resource management decisions. The purpose of this Technical Note is to highlight an improved modeling approach by combining the power of FloPy with the widely used Microsoft Excel software, to quickly utilize the power of an existing MODFLOW model.

Seismic Velocity Recovery in the Subsurface: Transient Damage and Groundwater Drainage Following the 2015 Gorkha Earthquake, Nepal

AbstractShallow earthquakes frequently disturb the hydrological and mechanical state of the subsurface, with consequences for hazard and water management. Transient post‐seismic hydrological behavior has been widely reported, suggesting that the recovery of material properties (relaxation) following ground shaking may impact groundwater fluctuations. However, the monitoring of seismic velocity variations associated with earthquake damage and hydrological variations are often done assuming that both effects are independent. In a field site prone to highly variable hydrological conditions, we disentangle the different forcing of the relative seismic velocity variations δv retrieved from a small dense seismic array in Nepal in the aftermath of the 2015 Mw 7.8 Gorkha earthquake. We successfully model transient damage effects by introducing a universal relaxation function that contains a unique maximum relaxation timescale for the main shock and the aftershocks, independent of the ground shaking levels. Next, we remove the modeled velocity from the raw data and test whether the corresponding residuals agree with a background hydrological behavior we inferred from a previously calibrated groundwater model. The fitting of the δv data with this model is improved when we introduce transient hydrological properties in the phase immediately following the main shock. This transient behavior, interpreted as an enhanced permeability in the shallow subsurface, lasts for ∼6 months and is shorter than the damage relaxation (∼1 yr). Thus, we demonstrate the capability of seismic interferometry to deconvolve transient hydrological properties after earthquakes from non‐linear mechanical recovery.

Improving Groundwater Model Calibration with Repeat Microgravity Measurements.

Groundwater-flow models depend on hydraulic head and flux observations for evaluation and calibration. A different type of observation-change in storage measured using repeat microgravity-can also be used for parameter estimation by simulating the expected change in gravity from a groundwater model and including the observation misfit in the objective function. The method is demonstrated using new software linked to MODFLOW input and output files and field data from the vicinity of the All American Canal in southeast California, USA. Over a 10-year period following lining of the previously highly permeable canal with concrete, gravity decreased by over 100 μGal (equivalent to about 2.5m of free-standing water) at some locations as seepage decreased and the remnant groundwater mound dissipated into the aquifer or was removed by groundwater pumping. Simulated gravity from a MODFLOW model closely matched observations, and repeat microgravity data proved useful for constraining both hydraulic conductivity and specific yield estimates. Specific yield estimated using the infinite-horizontal slab approximation agreed well with model-derived values, and the departure from the linear, flat-water-table approximation was small, less than 2%, despite relatively large and dynamic water-table slope. First-order second-moment parameter uncertainty analysis shows reduction in uncertainty for all hydraulic conductivity and specific yield parameter estimates with the addition of repeat microgravity data, as compared to drawdown data alone.

Hydrogeophysical model calibration and uncertainty analysis via full integration of PEST/PEST++ and COMSOL

Calibration of groundwater models is frequently hampered by limited availability of hydrogeological data. Non-intrusive geophysical methods are increasingly used to provide higher spatio-temporal resolution datasets for identification of hydrological processes and estimation of hydraulic properties. In groundwater model calibration performed through joint or coupled hydrogeophysical inversion, the hydrogeological datasets are supplemented with auxiliary geophysical data. In this work we propose a methodological approach to perform coupled inversion by integrating the calibration software PEST/PEST++ with COMSOL Multiphysics using MATLAB. The strategy provides multiple options for calibration and uncertainty analysis relevant for a broad range of environmental models. To illustrate the approach, we show a hydrogeophysical application in which electrical resistivity is used jointly with borehole data for the identification of seawater intrusion in a coastal aquifer.

Quantification of resilience metrics as affected by conservation agriculture at a watershed scale

It is suggested that conventional tillage operations exacerbate global environmental changes and affect the sustainability of our food production systems. Therefore, no-till has been introduced as one of the conservation agriculture practices to counteract these challenges. No-till has been adopted by a substantial number of farmers in major cropping regions; however, its resilience from large scale implementation has been overlooked. The majority of the studies have reported only a few aspects of the no-till practice (e.g., yield, soil properties, etc.), often with contradicting observations. To fill this gap, we present an approach that integrates long-term field experimental data and modeling to quantify resilience at a watershed scale. The study was conducted in the Kalamazoo River watershed located in Michigan, USA. Recharge, groundwater table, soil moisture, yield, and net return were used as resilience metrics. The DSSAT sequence crop model was developed for a corn-soybean-wheat rotation and calibrated using the yield and soil moisture data from a long-term (1993–2019) experiment for the conventional and the no-till treatment conducted within the study area. Soil moisture, recharge and yield were simulated, and the recharge was fed into a calibrated groundwater model to analyze changes in groundwater heads. The results illustrate clear evidence of higher recharge and net return under the no-till treatment, which were statistically significant for all crops at the watershed scale. Moreover, the no-till treatment consistently retained greater soil moisture than the conventional treatment, thereby helping to mitigate the impacts of droughts. The rise in groundwater table as affected by the adoption of no-till practices in this watershed has ranged between 0.1 and 0.5 m, depending on the underlying groundwater system, and has the potential to beneficially affect the aquifers and groundwater-dependent ecosystems. Therefore, the conservation agriculture could improve the overall resilience of the row crop system.

Utilizing Environmental Tracers to Reduce Groundwater Flow and Transport Model Parameter Uncertainties

AbstractNon‐uniqueness in groundwater model calibration is a primary source of uncertainty in groundwater flow and transport predictions. In this study, we investigate the ability of environmental tracer information to constrain groundwater model parameters. We utilize a pilot point calibration procedure conditioned to subsets of observed data including: liquid pressures, tritium (3H), chlorofluorocarbon‐12 (CFC‐12), and sulfur hexafluoride (SF6) concentrations; and groundwater apparent ages inferred from these environmental tracers, to quantify uncertainties in the heterogeneous permeability fields and infiltration rates of a steady‐state 2‐D synthetic aquifer and a transient 3‐D model of a field site located near Riverton, Wyoming (USA). To identify the relative data worth of each observation data type, the post‐calibration uncertainties of the optimal parameters for a given observation subset are compared to that from the full observation data set. Our results suggest that the calibration‐constrained permeability field uncertainties are largest when liquid pressures are used as the sole calibration data set. We find significant reduction in permeability uncertainty and increased predictive accuracy when the environmental tracer concentrations, rather than apparent groundwater ages, are used as calibration targets in the synthetic model. Calibration of the Riverton field site model using environmental tracer concentrations directly produces infiltration rate estimates with the lowest uncertainties, however; permeability field uncertainties remain similar between the environmental tracer concentration and apparent groundwater age calibration scenarios. This work provides insight on the data worth of environmental tracer information to calibrate groundwater models and highlights potential benefits of directly assimilating environmental tracer concentrations into model parameter estimation procedures.

GEOPOLIS software tools – Groundwater flow modeling at deep repository of liquid radioactive waste at Severny test site

This article discusses a three-dimensional groundwater flow model of a deep disposal facility at Severny test site. The three-dimensional model is a part of the certified software GEOPOLIS, based on the hydrogeological code GeRa (Geomigration of Radionuclides) serving as the calculation engine. This study describes the hydrogeological patterning of the groundwater flow model, as well as the results of calibration and verification of the model water heads with respect to the data of monitoring for more than 40 years of the deep repository exploitation. The article begins with a brief overview of the previously developed hydrogeological models of this object and continues with a description of the geological structure of the territory, and with a substantiation of the boundaries and parameters of the model. The results of groundwater flow modeling, model calibration, verification and estimation of discrepancy between the model results and monitoring data are shown. The comparison of the modeled and observed water heads in the stationary conditions (before the start of injection) and during operation of the deep repository allows making conclusion on the quality of calibration.

Automatic Calibration of Groundwater Models With Bias Correction and Data Filtering: Working With Drawdown Data

AbstractThe drawdown response to a hydraulic stress contains crucial information to characterize an aquifer. Modeling drawdowns is far easier than modeling heads because they are subject to homogeneous (zero) internal sink/sources, and boundary and initial conditions. The problem lies on the fact that drawdowns are not measured directly but derived from measurements of head fluctuations. Resulting drawdowns may suffer persistent inaccuracies in complex systems with uncertain long‐acting external stresses, so that they are affected not only by errors in head measurements, but also in estimates of the natural head evolution. This hinders the use of drawdowns in groundwater models, and forces modelers to employ absolute heads and soft information. In this context, we present a method to filter systematic errors in drawdown data during the calibration of a groundwater model. To do this, we introduce a bias correction term in a composite inverse problem that combines a natural head model with a drawdown model. Since these two models share the same parameters, a two‐stage iterative optimization algorithm is developed to jointly estimate the bias, natural trends, and parameters. The method is illustrated by a synthetic example in a heterogeneous aquifer. The example shows that the method converges to the best conditional estimate even when absolute head data is strongly biased. In the same example, we demonstrate that the use of biased absolute head data in the traditional inverse problem can also provide good fittings but, in this case, the bias leads to an incorrect estimation of the transmissivity field.

Relative importance of conceptual and computational errors when delineating saltwater intrusion from resistivity inverse models in heterogeneous coastal aquifers

In coastal areas, geological heterogeneity influences the geometry and characteristics of the freshwater-saltwater mixing zone. Geophysical electrical and electromagnetic methods are increasingly used to delineate the freshwater-saltwater mixing zone or for groundwater model calibration. However, in practical applications, it is common when assessing pore water salinity from resistivity models to disregard the spatial variability of the aquifer properties assuming pore salinity is the dominant control on bulk electrical conductivity. In this paper we use a coupled hydrogeophysical model to assess the discrepancies in recovered salinity caused by the geophysical inversion and the application of the chosen petrophysical model when heterogeneity is ignored. We find that errors made when neglecting heterogeneity during petrophysical transformation are equivalent or higher in magnitude than inversion errors, more so when an inappropriate petrophysical model is used, but exhibit different spatial distribution. The results show that their combined effect results in significant overestimation of the spreading of the mixing zone. The analysis provides new insights to the groundwater community to better evaluate the reliability of geophysically-derived information when delineating the freshwater-saltwater interface or when geophysics is used quantitatively as part of model scenario simulation, uncertainty analysis or optimisation strategies.

Groundwater budget deficit caused by drought and overexploitation

Abstract Groundwater is generally influenced by overexploitation and climatic stresses particularly in arid and semi-arid areas of the world. The present research was conducted to identify the relative contribution of drought and overexploitation to groundwater budget deficit in an unconfined aquifer system. In order to simulate groundwater, the simulated recharge from WetSpass-M model was applied in the MODFLOW model along with other required packages. Moreover, the groundwater budget deficit caused by stressors was quantified through the use of calibrated groundwater model predictions. In order to better understand how the stressors affect the groundwater deficit, the aquifer was divided into Clusters 1, 2, and 3. Locally, the results showed that the contribution of stressors to groundwater budget deficit was the highest in Cluster 1 due to the groundwater overexploitation and quick reaction of the groundwater level to the droughts. Overall, this research showed that both drought and overexploitation, with an average of 2.44 and 3.32 million cubic meters, respectively, played a significant role in groundwater storage deficit. Furthermore, the effect of groundwater overexploitation was approximately 36% more than droughts.