Heterogeneous Alluvial Aquifer Research Articles

Transport, dispersion, and degradation of nonpoint source contaminants during flood‐managed aquifer recharge

AbstractIn water‐stressed regions of the world, the inundation of working landscapes to replenish aquifers—known as flood‐managed aquifer recharge (flood‐MAR)—has become a valuable tool for sustainable groundwater management. Due to their diverse land use histories, however, many potential recharge sites host nonpoint source contaminants (such as salts, pesticides, and fertilizers) within the vadose zone that may flush to groundwater during recharge operations. To identify the controls on contaminant migration, we perform stochastic simulations of flood‐MAR through a heterogeneous alluvial aquifer and apply transient particle tracking to evaluate conservative and reactive contaminant transport over 80 years of recharge operations. With semi‐annual recharge events, the water table begins to rise 0.13–1.84 years after the first inundation event while solutes take much longer (11 to 80 years) to transit the 45‐m thick unsaturated zone. We derive a parametric expression for the ratio of celerity (or rate of pressure transmission) to velocity of the flood‐MAR wetting front and show that this simplified expression agrees with values calculated from heterogeneous model simulations. Slow solute velocities (0.25–1.75 m year−1) allow for significant contaminant removal through denitrification, but the contaminant plume experiences minimal dispersion or dilution over this time, reaching the water table as a sharp front. Our results suggest that minimizing groundwater velocity and maximizing groundwater celerity during flood‐MAR should optimize increases in water supply while limiting water quality degradation.

Adjoint models with non-Fickian reactive transport to identify pollutant sources in water

Pollutant source identification (PSI) has remained a research topic in environmental sciences for decades for hazardous chemicals, with the assumption that the transport of pollutants is Fickian. Pollutant transport in natural media, however, is usually non-Fickian, characterized by slow retention or preferential movement, and PSI models remained unknown for non-Fickian, reactive transport. To fill this knowledge gap, an adjoint fractional advection-dispersion-reaction equation (fADRE) was derived in this study to identify the source location of reactive pollutants undergoing non-Fickian diffusion with multiple phases in subsurface water with any dimension and boundary conditions. Six fundamental chemical reactions varying from radioactive decay to bimolecular reactions were considered, whose backward-in-time rate equations were simulated by particle tracking combined with backward non-Fickian diffusion. Applications showed that the adjoint fADRE identified the source location of reactive tracer transport in “homogeneous” sand columns repacked in our laboratory and a heterogeneous alluvial aquifer in an actual field site, while the Fickian-based, classical PSI model overestimated by four times the source location of reactive pollutants undergoing strong non-Fickian transport. To expand the applicability of the adjoint fADRE for PSI, we also discussed the impacts of the immobile pollutant phase and reactive kinetics on PSI, pollutant release history recovery, and usage of PSI for complex chemical reactions. The results showed that the fundamental chemical reactions considered in this study can affect PSI by adjusting the location and value of the peak of the probability density functions under conditions such as a relatively large decay constant or a space-dependent reaction rate, and the immobile pollutants can also affect PSI by shortening the source location and elongating the release time for pollutants usually detected only in the mobile phase. Therefore, non-Fickian dynamics and pollutant phases (which were missed by the adjoint ADRE), as well as reaction kinetics, must be captured by the adjoint model, such as the adjoint fADRE derived in this study, for PSI in real-world geological media.

Sustainability of Regional Groundwater Quality in Response to Managed Aquifer Recharge

AbstractGrowing demands on water supply worldwide have resulted in aquifer overdraft in many regions, especially in alluvial basins under intensive irrigation. This further leads to serious deterioration of groundwater quality. Managed aquifer recharge (MAR) has been shown to mitigate groundwater overdraft, but whether MAR can actually stabilize or reverse the ongoing declines in regional groundwater quality caused by non‐point sources has not been demonstrated. This study was intended to address the question by investigating impacts of different MAR strategies on regional groundwater quality. A geostatistical model was first used to characterize a heterogeneous alluvial aquifer system in a portion of the Tulare Lake Basin. Three‐dimensional numerical models were then employed to simulate groundwater flow and mass transport. Next, MAR strategies were applied in locations with different geological conditions or joint with different irrigation activities, and their performances were evaluated. Results demonstrate the potential of significant, long‐term benefits for regional groundwater quality by applying strategic, high‐intensity recharge operations on geologically favorable subregions. Siting MAR above the incised valley fill (IVF) deposit, a near‐surface paleochannel containing unusually coarse, high‐conductivity hydrofacies, leads to more extensive improvement in the groundwater quality in terms of salinity due to significant vertical flow and lateral outward flow from the IVF. Overall, decades would be required to alleviate groundwater quality concerns in the studied 189 km2 region. The simulations indicate that the deep concentrations remain below the secondary maximum contaminant level as the solute mass migrates downward with the prominent contribution from the attenuation via dispersion and matrix diffusion.

Estimating porosity distribution of a heterogeneous alluvial aquifer by GPU-accelerated 3D conditional stochastic inversion of common-offset GPR reflection data

Three-dimensional (3D) high-resolution and accurate characterization of the heterogeneity in an alluvial aquifer is critical to predicting the groundwater flow and contaminant transport. In this study, we extend the previous 2D conditional stochastic inversion method with respect to common-offset ground-penetrating radar (GPR) data to 3D for the estimation of 3D porosity distribution in a heterogeneous alluvial aquifer. First, 3D stochastic realizations of subsurface properties can be obtained using the fast Fourier transform moving average (FFT-MA) method, and the realizations are conditioned to borehole porosity measurements available in the 3D survey area, as well as to the geostatistical parameters derived from the borehole logs and the processed 3D GPR data. Next, the realizations are constantly modified and regenerated via a localized simulated annealing optimization strategy, which ultimately makes their corresponding synthetic data offer an acceptable fit to the GPR data. To accelerate the computational speed of 3D inversion, we design and implement the multi-GPUs multi-node parallelization of the 3D conditional stochastic inversion algorithm. The proposed 3D inversion algorithm is positively verified through the application to field GPR data collected at the Boise Hydrogeophysical Research Site near Boise, Idaho, USA. Our results indicate that the proposed 3D inversion method has the potential to effectively recover the fine-scale porosity distribution of a heterogeneous alluvial aquifer.

Backward Location and Travel Time Probabilities for Pollutants Moving in Three-Dimensional Aquifers: Governing Equations and Scale Effect

Backward probabilities have been used for decades to track hydrologic targets such as pollutants in water, but the convenient deviation and scale effect of backward probabilities remain unknown. This study derived backward probabilities for groundwater pollutants and evaluated their scale effect in heterogeneous aquifers. Three particle-moving methods, including the backward-in-time discrete random-walk (DRW), the backward-in-time continuous time random-walk (CTRW), and the particle mass balance, were proposed to derive the governing equation of backward location and travel time probabilities of contaminants. The resultant governing equations verified Kolmogorov’s backward equation and extended it to transient flow fields and aquifers with spatially varying porosity values. An improved backward-in-time random walk particle tracking technique was then applied to approximate the backward probabilities. Next, the scale effect of backward probabilities of contamination was analyzed quantitatively. Numerical results showed that the backward probabilities were sensitive to the vertical location and length of screened intervals in a three-dimensional heterogeneous alluvial aquifer, whereas the variation in borehole diameters did not influence the backward probabilities. The scale effect of backward probabilities was due to different flow paths reaching individual intervals under strong influences of subsurface hydrodynamics and heterogeneity distributions, even when the well screen was as short as ~2 m and surrounded by highly permeable sediments. Further analysis indicated that if the scale effect was ignored, significant errors may appear in applications of backward probabilities of groundwater contamination. This study, therefore, provides convenient methods to build backward probability models and sheds light on applications relying on backward probabilities with a scale effect.

Potential effects on groundwater quality associated with infiltrating stormwater through dry wells for aquifer recharge

Dry wells (gravity-fed infiltration wells) have frequently been used to recharge aquifers with stormwater, especially in urban areas, as well as manage flood risk and reduce surface water body contamination from stormwater pollutants. However, only limited assessment of their potential adverse impacts on groundwater quality exists. Dry well recharge can bypass significant portions of the filtering-capacity of the vadose zone. Stormwater and groundwater monitoring data and analysis of transport of a wide range of historic and current-use stormwater chemicals of concern is lacking. To address these gaps, two dry wells were constructed with vegetated and structural pretreatment features to assess the likelihood of stormwater contaminants reaching the aquifer. We monitored, assessed, and compared the presence of contaminants in stormwater to water quality in the vadose zone and shallow groundwater after it passed through the dry well. The dry wells were installed at a suburban residential and at a suburban commercial site. The selected sites were overlying a regional, unconsolidated, and highly heterogeneous alluvial aquifer system. Stormwater, vadose zone, and groundwater samples were collected during five storms and analyzed for over 200 contaminants of concern. Relatively few contaminants were detected in stormwater, generally at low concentrations. Prior to stormwater entering the dry well, 50–65% of contaminants were removed by vegetated pretreatment. In groundwater, metals such as aluminum and iron were detected at similar concentrations in both upgradient and downgradient wells, suggesting the source of these metals was not dry well effluent. Naturally occurring metals such as chromium and arsenic were not detected in stormwater but were found at elevated concentrations in groundwater. A modeling assessment suggests that the travel time of metals and hydrophobic organic contaminants to the water table at these sites ranges from years to centuries, whereas water soluble pesticides would likely reach the water table within days to months. The modeling assessment also showed that more vulnerable sites with higher fraction of alluvial sands would have much shorter contaminant travel times. However, none of the contaminants assessed reached concentrations that pose a risk to human health across the scenarios considered. No evidence was found, either through direct measurements or vadose zone modeling, that contaminants present in suburban stormwater degraded or would degrade groundwater quality at the studied sites and site conditions. Future work is needed to address emerging contaminants of concern.

Coupling hydrogeophysics with hydrodynamic modelling to infer subsurface hydraulic architecture of an alluvial floodplain

ABSTRACTThis paper underscores the importance of spatially dense geophysical data sets for making informed decisions in water management strategies. Such decisions may require understanding how site‐specific subsurface architecture – especially hydraulic connectivity – impacts the response of a shallow aquifer to anthropogenic hydrologic disturbances (e.g. over‐pumping of a shallow aquifer). At a 0.2‐km2 alluvial floodplain site characterized by thick clay over fine sand to gravel and shale bedrock in the subtropical, sub‐humid belt of the Gulf Coast of the United States, we image an asymmetrically shaped, compartmentalized, sand‐dominated channel‐belt using electrical resistivity tomography and 31 time‐domain electromagnetic soundings probing to depths of ∼40 m and ∼90 m, respectively. Lithological interpretation and a hydrological model are developed based on the geophysical data and nearby sediment cores, where the resistivity of the groundwater is 9.1 Ωm. In a modelling scenario wherein the compartmentalized sand channel‐belt starts out dry (i.e. an over‐pumped shallow aquifer), we simulate 26 weeks of infiltration due to flooding of the surface. Preferential filling of the channel‐belt occurs through its sides rather than from above, generating a new understanding of the hydraulic connectivity around and into asymmetrically shaped, sand‐dominated channel‐belts. This insight can inform decisions about the optimal placement of shallow water wells in a heterogeneous alluvial floodplain aquifer system and also highlights the dangers of over‐pumping.

Imaging the hydraulic properties of a contaminated alluvial aquifer perturbed with periodic signals

Imaging characterization of a heterogeneous alluvial aquifer at a decametric scale is presented. The characterization relies on responses to oscillatory pumping tests led in two different wells and at two different periods of oscillation (5 and 10 min). These specific oscillatory responses are extracted from the hydraulic pressure values, measured in 13 boreholes during the pumping tests, through a data processing procedure. Then, a deterministic inversion process, led with a two-dimensional hydraulic properties model, aims to reproduce this set of oscillatory responses, in a frequency domain, by optimizing the distribution of the property values in the model. Two inversion processes are led separately with each set of responses corresponding to the two periods of oscillation used, and one joint inversion is led with the set of all responses together. The inversion results image the heterogeneities in the distribution of the field properties. The results suggest that longer periods of oscillation (in this case 10 min) permit characterization of larger areas around the pumping wells. In contrast, shorter periods (in this case 5 min) propagate more readily in the conductive zones near the pump and provide less information for the less conductive areas. Therefore, it appeared that performing a joint inversion, combining the information carried by the responses to both periods, provided more information on the heterogeneous distribution of the field properties and resulted in better constrained property maps than the ones obtained from separate inversions.

Quaternary Alluvial Aquifer Study Using Integrated Geophysical Approach in Zaghouan Plain (Northeastern Tunisia)

The study area Zaghouan Plain (northeastern Tunisia) is unique for its strategic Quaternary aquifer, which is, in a great proportion, composed of sand lenses that vary in thicknesses and forms. Conducting an integrative approach based on regional mapping and local studies helped provide precise calibration to identify convenient areas for hydrogeological explorations, which are focused on heterogeneous alluvial and fluvio-deltaic aquifers. To enhance the location and the extension of low density related to sand lenses porous aquifers, a regional gravity analysis supported by focused seismic interpretation was performed. Thus, advanced analyses were conducted in the Zaghouan Plain through the implementation of a regional–residual separation, a radially averaged power spectrum and tilt/TDX filter density models. During the process, two major negative gravity anomalies, associated with developed lenses system between Zaghouan Mountain and Henchir Bouslim anticline, were revealed. The anomalies differed in direction from E–W to NW–SE. They also varied widthwise from 1000 to 3000 m and depthwise from 1500 to 2000 m. Hence, with the aim of highlighting the detailed geometry of sand lenses and locating the most interesting zone for future hydrogeological exploration and/or recharge process, 1D time-domain electromagnetic and 2D electrical resistivity tomography studies were conducted and various techniques were applied in a local area situated in the main negative gravity anomaly.

Stochastic Assessment of Nonpoint Source Contamination: Joint Impact of Aquifer Heterogeneity and Well Characteristics on Management Metrics

AbstractNonpoint source (NPS) groundwater contamination in sedimentary basin aquifers with overlying agricultural activities increasingly threatens groundwater supplies. The role of aquifer heterogeneity has not been well understood in the assessment of NPS and in linking pollution sources to impacts in water supply wells. A typical well taps into and mixes groundwater varying in age by decades or even centuries. This study investigates the joint impact of aquifer heterogeneity and pumping well characteristics (well depth, pumping rate, and screen length) on expected values of and uncertainty about key management metrics: (1) travel time of a NPS contaminant to a production well, (2) spatiotemporal characteristics of the source area, and (3) contaminant compliance at production wells. A stochastic approach is employed using Monte Carlo simulation of flow and nonreactive transport in 3‐D highly heterogeneous alluvial aquifer systems. Well design is shown to dominate the distribution of travel time mixing and the overall location and spread of the source area. Larger extraction intensity and closer proximity to the land surface are shown to significantly suppress effects of large aquifer heterogeneity on uncertainty on all metrics. Deep wells and wells with lower pumping rates have more uncertain source areas. Long‐term NPS contaminants with high source concentration (>10 times compliance level) will be exceeded in most wells within decades, while low‐intensity source concentrations (2–4 times compliance level, typical for nitrate and salinity) have large uncertainty about time to exceedance suggesting wide variability among a set of wells with similar well design subject to the same NPS pollution.

Differentiating anthropogenic and natural sources of uranium by geochemical fingerprinting of groundwater at the Homestake uranium mill, Milan, New Mexico, USA

A multiparameter geochemical-isotopic fingerprinting approach was used to differentiate anthropogenic and natural signatures of uranium contamination near the Homestake uranium mill site (Site), near Milan, New Mexico, USA. The Site consists of two tailings piles from milling operations and groundwater contamination from these tailings has been noted. The Site lies within the lower San Mateo Creek Basin and has multiple regional sources of uranium contamination from mining and mill operations. The Site is underlain by a heterogeneous alluvial aquifer, which is in turn underlain by basement rock of the Chinle Group aquifer and the underlying San Andres-Glorieta Formation aquifer. To help decipher signatures, several statistical approaches were used including principal component analysis, non-metric multidimensional scaling, and cluster analysis. Piper diagrams indicate two end-member water types at the Site, sulfate–Na–K generally in the Chinle Group aquifer and sulfate–Ca generally in the alluvial aquifer. There are wells from both aquifers that plot between the two end members. Uranium concentrations from the Site fall into three broad categories: less than the drinking water standard of 30 µg/L (n = 3), from 30 to 100 µg/L (n = 9), and greater than 100 µg/L (n = 8). Component loadings in a principal component analysis are highest for uranium isotopes, uranium, molybdenum, chloride, sodium, 228radium, and gross alpha–beta, which affect the similarities or differences among wells sampled. Results suggest that several alluvial wells north of the Site have groundwater with anthropogenic fingerprints from regional sources related to upgradient mining. Well water with higher uranium concentrations has uranium activity ratios close to 1, which is indicative of mining or milling signatures. These same wells have elevated radon activities. This information can be used to inform Site managers regarding the source of water related to uranium at the Site and provide an approach for geochemical fingerprinting.

Using solute and heat tracers for aquifer characterization in a strongly heterogeneous alluvial aquifer

A test using Rhodamine WT and heat as tracers, conducted over a 78 day period in a strongly heterogeneous alluvial aquifer, was used to evaluate the utility of the combined observation dataset for aquifer characterization. A highly parameterized model was inverted, with concentration and temperature time-series as calibration targets. Groundwater heads recorded during the experiment were boundary dependent and were ignored during the inversion process. The inverted model produced a high resolution depiction of the hydraulic conductivity and porosity fields. Statistical properties of these fields are in very good agreement with estimates from previous studies at the site. Spatially distributed sensitivity analysis suggests that both solute and heat transport were most sensitive to the hydraulic conductivity and porosity fields and less sensitive to dispersivity and thermal distribution factor, with sensitivity to porosity greatly reducing outside the monitored area. The issues of model over-parameterization and non-uniqueness are addressed through identifiability analysis. Longitudinal dispersivity and thermal distribution factor are highly identifiable, however spatially distributed parameters are only identifiable near the injection point. Temperature related density effects became observable for both heat and solute, as the temperature anomaly increased above 12 degrees centigrade, and affected down gradient propagation. Finally we demonstrate that high frequency and spatially dense temperature data cannot inform a dual porosity model in the absence of frequent solute concentration measurements.

Reactive transport in the complex heterogeneous alluvial aquifer of Fortymile Wash, Nevada

Yucca Mountain, Nevada, had been extensively investigated as a potential deep geologic repository for storing high-level nuclear wastes. Previous field investigations of stratified alluvial aquifer downstream of the site revealed that there is a hierarchy of sedimentary facies types. There is a corresponding log conductivity and reactive surface area subpopulations within each facies at each scale of sedimentary architecture. Here we use a Lagrangian-based transport model in order to analyze radionuclide dispersion in the saturated alluvium of Fortymile Wash, Nevada. First, we validate the Lagrangian model using high-resolution flow and reactive transport simulations. Then, we used the validated model to investigate how each scale of sedimentary architecture may affect long-term radionuclide transport at Yucca Mountain. Results show that the reactive solute dispersion developed by the Lagrangian model matches the ensemble average of numerical simulations well. The link between the alluvium spatial variability and reactive solute dispersion at different spatiotemporal scales is demonstrated using the Lagrangian model. The longitudinal dispersivity of the reactive plume can be on the order of hundreds to thousands of meters, and it may not reach its asymptotic value even after 10,000 years of travel time and 2–3 km of travel distance.

Validity of flowmeter data in heterogeneous alluvial aquifers

Numerical simulations are performed to evaluate the impact of medium-scale sedimentary architecture and small-scale heterogeneity on the validity of the borehole flowmeter test, a widely used method for measuring hydraulic conductivity (K) at the scale required for detailed groundwater flow and solute transport simulations. Reference data from synthetic K fields representing the range of structures and small-scale heterogeneity typically observed in alluvial systems are compared with estimated values from numerical simulations of flowmeter tests. Systematic errors inherent in the flowmeter K estimates are significant when the reference K field structure deviates from the hypothetical perfectly stratified conceptual model at the basis of the interpretation method of flowmeter tests. Because of these errors, the true variability of the K field is underestimated and the distributions of the reference K data and log-transformed spatial increments are also misconstrued. The presented numerical analysis shows that the validity of flowmeter based K data depends on measureable parameters defining the architecture of the hydrofacies, the conductivity contrasts between the hydrofacies and the sub-facies-scale K variability. A preliminary geological characterization is therefore essential for evaluating the optimal approach for accurate K field characterization.

A partial exponential lumped parameter model to evaluate groundwater age distributions and nitrate trends in long-screened wells

A partial exponential lumped parameter model (PEM) was derived to determine age distributions and nitrate trends in long-screened production wells. The PEM can simulate age distributions for wells screened over any finite interval of an aquifer that has an exponential distribution of age with depth. The PEM has 3 parameters – the ratio of saturated thickness to the top and bottom of the screen and mean age, but these can be reduced to 1 parameter (mean age) by using well construction information and estimates of the saturated thickness. The PEM was tested with data from 30 production wells in a heterogeneous alluvial fan aquifer in California, USA. Well construction data were used to guide parameterization of a PEM for each well and mean age was calibrated to measured environmental tracer data (3H, 3He, CFC-113, and 14C). Results were compared to age distributions generated for individual wells using advective particle tracking models (PTMs). Age distributions from PTMs were more complex than PEM distributions, but PEMs provided better fits to tracer data, partly because the PTMs did not simulate 14C accurately in wells that captured varying amounts of old groundwater recharged at lower rates prior to groundwater development and irrigation. Nitrate trends were simulated independently of the calibration process and the PEM provided good fits for at least 11 of 24 wells. This work shows that the PEM, and lumped parameter models (LPMs) in general, can often identify critical features of the age distributions in wells that are needed to explain observed tracer data and nonpoint source contaminant trends, even in systems where aquifer heterogeneity and water-use complicate distributions of age. While accurate PTMs are preferable for understanding and predicting aquifer-scale responses to water use and contaminant transport, LPMs can be sensitive to local conditions near individual wells that may be inaccurately represented or missing in an aquifer-scale flow model.

A lithofacies approach for modeling non‐Fickian solute transport in a heterogeneous alluvial aquifer

AbstractStochastic realizations of lithofacies assemblage based on lithological data from a relatively small number of boreholes were used to simulate solute transport at the well‐known Macrodispersion Experiment (MADE) site in Mississippi (USA). With sharp vertical contrasts and lateral connectivity explicitly accounted for in the corresponding hydraulic conductivity fields, experimental results from a large‐scale tracer experiment were adequately reproduced with a relatively simple model based on advection and local dispersion. The geologically based model of physical heterogeneity shows that one well‐interconnected lithofacies, with a significantly higher hydraulic conductivity and accounting for 12% of the total aquifer volume, may be responsible for the observed non‐Fickian transport behavior indicated by the asymmetric shape of the plumes and by variations of the dispersion rate in both space and time. This analysis provides a lithological basis to the hypothesis that transport at MADE site is controlled by a network of high‐conductivity sediments embedded in a less permeable matrix. It also explains the calibrated value of the ratio of mobile to total porosities used in previous modeling studies based on the dual‐domain mass transfer approach. The results of this study underscore the importance of geologically plausible conceptualizations of the subsurface for making accurate predictions of the fate and transport of contaminants in highly heterogeneous aquifers. These conceptualizations may be developed through integration of raw geological data with expert knowledge, interpretation, and appropriate geostatistical methods.

The propagation of complex flood-induced head wavefronts through a heterogeneous alluvial aquifer and its applicability in groundwater flood risk management

Summary The extraordinary rise of piezometric surface in aquifers may induce an unexpected interaction of groundwater with anthropogenic elements known as groundwater inundation phenomena. This could result in several damage processes, including building foundation destabilization, groundwater infiltration and pollutant remobilization, which are responsible for considerable economic losses worldwide. To improve our knowledge concerning flood-induced head wavefront propagation and groundwater inundation phenomena, the reconstructed kinematics of the head wavefronts obtained from a calibrated groundwater model simulating 12 ordinary flood events have been statistically analyzed. The correlation between kinematic variables of a flood-induced head wavefront, flood event characteristics and aquifer properties was the basis for the study of the aquifer response to river flood events along four different trajectories, i.e. the identification of river level rise rate, absolute–relative height of the maximum rise stage, hydraulic parameter variability, river–aquifer exchange rates, pre-event state of the aquifer and distance to the aquifer boundary as the key-influencing factors in groundwater inundation. Finally, model and scenario results were used to develop two synthetic models for groundwater inundation as novel decision-support tools for assessing both ordinary and extraordinary flood events in groundwater flood risk management and in urban development planning. These synthetic models gives the technical criteria to predict extraordinary rise of piezometric surface and therefore allows increasing the confidence of stakeholders in the risk assessment and improving the subsurface infrastructure design to mitigate damage processes derived from groundwater interaction.

EnKF coupled with groundwater flow moment equations applied to Lauswiesen aquifer, Germany

Summary We describe a field application of a new data assimilation method recently proposed by Panzeri et al. (2013, 2014). The method couples a modified Ensemble Kalman Filter (EnKF) algorithm with stochastic moment equations (MEs) governing space–time variations of (theoretical ensemble) mean and covariance values of groundwater flow state variables (hydraulic heads and fluxes). Whereas traditional EnKF entails Monte Carlo (MC) simulations and suffers from inbreeding, our approach obviates both. Synthetic case studies have shown the ME-based approach to be computationally efficient and accurate when compared to MC-based results. Here we use our ME-based method to assimilate drawdown data recorded during cross-hole pumping tests in the heterogeneous alluvial Lauswiesen aquifer near Tubingen, Germany. Our results include an estimate of log transmissivity distribution throughout the aquifer and corresponding measures of estimation error. We validate our calibrated model by using it to predict drawdowns recorded during another pumping test at the site and compare its performance with that of standard MC-based EnKF.

Predicting flow and transport in highly heterogeneous alluvial aquifers

AbstractSuccessful prediction of groundwater flow and solute transport through highly heterogeneous aquifers has remained elusive due to the limitations of methods to characterize hydraulic conductivity (K) and generate realistic stochastic fields from such data. As a result, many studies have suggested that the classical advective‐dispersive equation (ADE) cannot reproduce such transport behavior. Here we demonstrate that when high‐resolution K data are used with a fractal stochastic method that produces K fields with adequate connectivity, the classical ADE can accurately predict solute transport at the macrodispersion experiment site in Mississippi. This development provides great promise to accurately predict contaminant plume migration, design more effective remediation schemes, and reduce environmental risks.

Relative rates of solute and pressure propagation into heterogeneous alluvial aquifers following river flow events

Conventional theory for homogeneous aquifers states that pressure propagates more rapidly into aquifers than solutes following river stage rise. We demonstrate through numerical simulations of two-dimensional aquifer slices that the relative timing of pressure and solute responses in alluvial aquifers is a function of subsurface structures. Two generic conceptual models of heterogeneity are investigated, a vertical clogging layer and a horizontal sand string. Independent of the conceptual model, the hydraulic conductivity contrast is the primary controlling variable on the rates of pressure and solute transport from a river to an observation point. Conceptual models are compared using metrics for pressure and solute travel time that represent propagation of 50% change in each variable from river to observation point. While not possible in a homogeneous system, a solute travel time less than a pressure travel time can occur in the presence of both types of heterogeneity, and indicates that heterogeneity is controlling propagation from the river to the aquifer. Less than one order of magnitude contrast in hydraulic conductivities is sufficient to create a travel time ratio less than one. Contrasts of this magnitude are often exceeded in alluvial environments and thus simultaneous measurement of solute and pressure has the potential to constrain estimates of exchange flux in a way not possible with pressure measurements alone. In general, flux estimates derived from solute travel times provide more accurate estimates than those derived from pressure responses in heterogeneous environments. The magnitude of error in estimates derived from pressure responses is proportional to the hydraulic conductivity contrast. Travel times calculated from time series pressure and EC data collected in the Mitchell River in northern Australia are used to demonstrate application of this combined approach.