La gestion durable des aquifères est une problématique grandissante depuis la fin du 20ème siècle. L'exploitation d’une ressource en eau souterraine est qualifiée de durable lorsque la capture des flux environnementaux est considérée comme acceptable sur le long terme. La modélisation hydrodynamique s'impose comme un outil indispensable pour remplacer une gestion réactive par une approche anticipative. Les paramètres hydrodynamiques qui caractérisent un aquifère et contrôlent les variables de sorties des modèles hydrodynamiques sont souvent mal connus. L’estimation de ces paramètres par la modélisation inverse souffre de la non-unicité de la solution optimale. Une approche simplifiée pour la quantification des incertitudes (analyse linéaire) est présentée comme une alternative pragmatique à des méthodes stochastiques inapplicables pour des modèles opérationnels. A partir de la réalisation d’une station expérimentale pilote, différentes méthodes (parfois complémentaires) ont été évaluées pour contraindre la recharge météorique et les propriétés hydrauliques d’un aquifère afin de réduire l’incertitude prédictive. La réalisation d’un modèle vertical couplé sol-surface a permis de démontrer que, dans le contexte étudié, la tension matricielle apporte suffisamment d’informations afin de contraindre la recharge prédite. Une interprétation conjointe d’un essai de nappe libre et des fluctuations piézométriques a permis une estimation intégrée de la recharge et des paramètres hydrodynamiques de la nappe libre. Ce travail de thèse a ainsi permis (i) de démontrer l’intérêt de disposer de méthodes algorithmiques pour la calibration et la quantification des incertitudes paramétriques pour un modèle hydrodynamique de gestion ; (ii) de mener une réflexion méthodologique sur l’utilisation de méthodes existantes afin d’apporter de l’information complémentaire fiable sur les paramètres hydrodynamiques ainsi que sur la recharge météorique. Ce travail offre des perspectives quant à la mise en place d’un réseau de suivi complet à l’échelle d’un bassin hydrogéologique. / The sustainable management of aquifers is a growing problem since the end of the 20th century. For groundwater withdrawals to be considered as sustainable, the capture of environmental flow should remain acceptable over a long-term period. Groundwater modeling is an essential tool to move from a reactive management to an anticipatory approach. Hydrodynamic parameters characterizing the aquifers are often poorly constrained by prior information or history matching. The estimation of these parameters by inverse modeling suffers from the non-uniqueness of the solution. This is an issue when predicted values by groundwater model are used to define legal frameworks. A simplified approach for the quantification of uncertainties (linear analysis) is presented as a pragmatic alternative to stochastic methods that cannot be applied to operational groundwater management models. The implementation of a pilot experimental station brings possibility to evaluate different approaches for the estimate of groundwater recharge and hydrodynamic parameters estimation in order to reduce the uncertainty of groundwater management models. A 1D coupled soil-surface model was used to demonstrate that, in the studied context, matrix potential measurements alone appear as sufficient to constrain coupled model-based estimates of recharge. In addition, a joint interpretation of an unconfined aquifer-test and water table fluctuations has been conducted. Reliable estimates of groundwater recharge can be obtained from water level records when considering long recharge events and a consistent value of drainable porosity. This thesis highlights (i) the necessity to use algorithmic methods for parameters estimation and uncertainty quantification for a groundwater management model; (ii) the interest of different methods to collect reliable hydrodynamic parameters and groundwater recharge estimation. This work can be used to support a monitoring network for parameters estimation at a basin scale.

Hydraulic fracturing is used to extract oil and natural gas from low permeability formations. The potential of fluids migrating from depth through adjacent wellbores and through the production wellbore was investigated using statistical modeling and predic-tive classifiers. The probability of a hydraulic fracturing well becoming hydraulically connected to an adjacent well in the Marcellus shale of New York was determined to be between 0.00% and 3.45% at the time of the study. This means that the chance of an in-duced fracture from hydraulic fracturing intersecting an existing well is highly dependent on the area of increased permeability caused by fracturing. The chance of intersecting an existing well does not mean that fluid will flow upwards; for upward migration to occur, a pathway must exist and a pressure gradient is required to drive flow, with the exception of gas flow caused by buoyancy. Predictive classifiers were employed on a dataset of wells in Alberta Canada to identify well characteristics most associated to fluid migration along the production well. The models, specifically a random forest, were able to identify pathways better than random guessing with 78% of wells in the data set identified cor-rectly. Magnetic resonance imaging (MRI) was used to visualize and quantify contami-nant transport in a soil column using a full body scanner. T1 quantification was used to determine the concentration of a contaminant surrogate in the form of Magnevist, an MRI contrast agent. Imaging showed a strong impact from density driven convection when the density difference between the two fluids was small (0.3%). MRI also identified a buildup of contrast agent concentration at the interface between a low permeability ground silica and higher permeability AFS 50-70 testing sand when density driven con-vection was eliminated.

Stakeholder participation is a foundation of good water governance. Good groundwater governance typically involves the co-production of knowledge about the groundwater system. Models provide a vehicle for producing this knowledge, as well as a boundary object around which scientists and stakeholders can convene the co-production process. Through co-production, stakeholders and scientific experts can engage in exchanges that create system knowledge not otherwise achievable. The process involves one-way transfer of information, active two-way conversations, and integration of multiple kinds of knowledge into shared understanding. In the Upper Santa Cruz River basin in Arizona, USA, the University of Arizona Water Resources Research Center (WRRC) convened a project aimed at providing scientific underpinnings for groundwater planning and management. This project, entitled Groundwater, Climate, and Stakeholder Engagement, serves as a case study employing the first two stages of knowledge co-production using a hydrological model. Through an iterative process that included two-way communication, stakeholders provided critical input to hydrologic modeling analyses. Acting as a bridging organization, the WRRC facilitated a co-production process, involving location-specific and transferability workshops, which resulted in new knowledge and capacity for applying the model to novel problems.

No description available.

Bwaise III in Kampala, Uganda is a densely populated informal settlement with a shallow groundwater table and inadequate basic services. High risk of groundwater contamination will bring health problem to the local residents. In this study, a large dimension (300 m in length) 2D model was developed to depict the hydrogeological condition and to examine the response to different rainfall infiltration rate on the groundwater table. The boundary condition of the drainage system plays an important role in modeling the groundwater flow. The simulation results show that water in the drain will flow into the aquifer when the drain is full, otherwise the drain will act as a sink for ground water. Advective transport of phosphorus results in no pollutants reaching or percolating into the drain. The integration of phosphorous concentra-tion flowing out of Domain 3 (pollutant inlet) corresponds to the infiltration rate and the plume moves faster during the wet season which brings in more phosphorous compared with the dry season. With sorption, all the phosphorus was adsorbed within the top soil. A simplified 3D model was set up to illustrate the flow field. Additional simulation can be undertaken within this 3D frame for more realistic calculation and consistent prediction.

No description available.

Karst aquifers are dynamic hydrologic systems which are sensitive to short-term recharge events (storms) and heterogeneous recharge characteristics (point recharge at sinks, irregular soil thicknesses). These aquifers are highly productive yet also vulnerable to contamination, in large part because the conduit network is a significant unknown for predicting karst flow paths. To address these uncertainties, two adjacent karst springs, Tippery Spring and Near Tippery Spring, were monitored to better understand flow and source mixing characteristics. The two springs in central Pennsylvania’s Nittany Valley have similar discharges and are only 65 meters apart, yet they show unique behaviors in terms of water chemistry and discharge response to storms. First examined for flow characterization in 1971 by Shuster and White, the springs were analyzed in this study using high-resolution logging and new tracers such as rare earth element (REEs) and Ca/Zr ratios. This research contributes to the field of karst hydrology through innovative water sampling and monitoring techniques to investigate karst recharge and flow behavior along with conduit flow models incorporating multiple calibration target datasets such as water temperature and dye tracing. Stable isotope signatures (δD & δ18O) of storm water samples at the two springs varied based on storm intensity, but also due to their unique recharge behaviors. Increased spring discharge preceded the arrival of storm water as conduits were purged of pre-storm water, indicated by no change in isotopic composition on the rising limb. The isotopic signature then became progressively more enriched at both springs, indicating storm water recharge. At Tippery, this enrichment began around peak flow, sooner than at Near Tippery where enrichment began during the descending limb. Thus, isotopes indicated a stronger surface connection at Tippery Spring. Storm intensity also affected the relative contribution of recharging water reaching both springs, with a larger storm producing a larger recharge signature compared to a smaller storm. At Tippery Spring, for a short time the majority of emerging water was storm water, which may indicate a reversal in water exchange between the conduits and the surrounding matrix, an important consideration in karst contaminant transport. Two natural tracers were applied in new ways for this study: Ca/Zr ratios and REE patterns. Both tracers provided additional information about flow paths and recharge sources as they varied during the storm hydrograph. Ca/Zr ratios changed in timing and intensity with storm intensity, and both springs exhibited a decline in Ca/Zr ratios as calcium-rich carbonate matrix water was displaced by zirconium-rich storm recharge water from sinking streams off the clastic upland ridges. Being a storm water arrival indicator in clastic-ridge-fed Valley and Ridge springs, this relationship made Ca/Zr ratios a useful substitute for stable water isotopes while also providing information on source area. In response to storm water recharge, REE concentrations increased with the arrival of storm water. The timing and magnitude of concentration increases were influenced both by the degree of surface connectivity intrinsic to each spring and the intensity of the recharge event. Elevated REE concentrations persisted after other parameters recovered to pre-storm levels, suggesting water which has interacted with either the local carbonate matrix or the upland siliciclastics. These slower flow paths recharging the two springs were not apparent from other geochemical parameters. This study illustrated the relationships among multiple tracers to understand source waters in different periods of storm hydrographs. A flow and transport model using the Finite Element Subsurface Flow Model (FEFLOW) was calibrated using quantitative dye trace and high resolution temperature data to simulate the connection between a sinking stream and Tippery Spring. Dye was injected at the sink and monitored at the spring while temperature data was collected using loggers at both the sink and the spring. FEFLOW was used to simulate the connection between sink and spring through varying conduit geometries, sink and spring discharges, conduit conductivity, conduit cross-sectional area, matrix transmissivity, matrix porosity, and dispersivity. Single conduit models reproduced larger peak and recession concentrations than observed. A forked conduit model diverted flow from the main conduit, reducing the concentration of dye reaching the spring, provided a better match. Latin Hypercube sensitivity analysis indicated that dye concentration breakthrough curves were most sensitive to conduit conductivity and less sensitive to other model parameters. Temperature data from high-resolution loggers at the sink and spring were then incorporated into the model scenarios to reproduce seasonal spring temperature using the conduit configuration fit to the dye trace. Simulated temperature signals at the spring were sensitive to parameters in addition to conduit conductivity, most notably matrix transmissivity and inflow rates at the sink. The dual approach to karst model calibration using a temperature model set up from an initial dye trace results in greater model confidence due to a limited possible range in conduit conductivity. This study improved conceptual and numerical models for karst by examining how data from storm events and tracers can be used to better understand recharge and flow paths. / Geoscience

Streams and rivers are a vital part of our ecosystem. They are imperiled by human ecological activities such as urbanization, industrialization, and agriculture which discharge excess nitrate and other pollutants into our waterways. Here, this dissertation seeks to understand the physical and biogeochemical processes which attenuate pollutants in stream corridors. The focus is hyporheic zones which form the interface between surface water and groundwater below and adjacent to stream channels, and riparian zones which form the interface between channels and adjacent uplands, both of which can attenuate pollutants. In this context, soil-pipes can dominate subsurface hydraulics. This research first employed MODFLOW and MT3D-USGS to model transient hyporheic hydraulics and nitrate transport in a length of riparian/riverbank soil to probe the effects of soil pipes on hydraulics and denitrification due to peak flow events in the channel. Findings showed that inserting just one soil pipe 1.5 m in length caused a ~75% increase in both hyporheic exchange and denitrification. A rough upscaling showed soil pipes could remove up to ~3% of nitrate along a 1-km reach. Next, the ability of soil pipes to bypass the often championed ability of riparian buffers to remove nitrate migrating from uplands to the channel was evaluated. This effort also employed MODFLOW and MT3D-USGS to simulated hydraulics and nitrate removal along a length of riparian soil. Findings showed that soil pipes increased flow of nitrate to the banks by five orders of magnitude in some cases. We posited a non-dimension parameter which governs when nitrate bypass is significant. In addition to soil pipes, dune bedforms can also enhance hyporheic exchange, primarily in the stream/riverbed. Again employing MODFLOW but now pairing with the transport code SEAM3D to simulate microbially-mediated aerobic metabolism of dissolved organic carbon and dissolved oxygen, the combined effects of dune translation and microbial growth and death were explored. Major findings include that neglecting microbial growth can lead to inaccurate modeling of biogeochemistry, and that aerobic metabolism increased with celerity. The results herein bolster knowledge of natural pollutant attenuation in stream and river corridors, and have implications for pollutant mitigation strategy and stream credit allocation. / Doctor of Philosophy / Streams are a vital part of our ecosystem. They are imperiled by human ecological activities such as urbanization, industrialization, and agriculture which discharge nitrate and other pollutants into our waterways. Here, this dissertation seeks to understand the physical and biological processes which attenuate pollutants. The hyporheic zone is the interface between surface water and groundwater below the bed and adjacent to stream banks, and can attenuate pollutants. Transient peak flow events such as a storm or snow melt raise the stream water levels, causing the water pressure in the stream channel to temporarily outweigh the water pressure in the soil pore spaces adjacent to the stream channel. This drives water into the banks subjecting it to pollutant attenuation processes. Soil pipes (long cylindrical void spaces created by decayed plant roots) are prevalent along stream banks, and they dominate subsurface hydraulics. This dissertation implemented a numerical study on a chunk of riparian soil to probe the effects of soil pipes on hydraulics and denitrification. Findings showed that inserting just one – 1.5 m soil pipe caused a ~75% increase in both water flow volume into the bank and nitrate removal. Riparian buffers are the vegetated strips adjacent to stream channels and have long been championed as stalwarts of pollutant removal. Soil pipes undermine this by acting as a bypass mechanism. A numerical study was again performed on a chunk of riparian soil to quantify the effects soil pipes on riparian bypass of nitrate. Findings showed that soil pipes increased flow of nitrate to the banks by five orders of magnitude in some cases. This means that a buffer enhancement strip with fine roots that prevent the formation of soil pipes should be installed along riparian buffers. In addition to soil pipes, dune bedforms can increase flowrate of water into the hyporheic zone. This dissertation modeled the combined effects of dune translation and microbial growth and death. Major findings include that neglecting microbial growth can lead to inaccurate modeling of biogeochemistry, and that biodegradation increases with increased dune velocity. The results herein bolster knowledge on natural pollutant attenuation in streams, and have implications in terms of pollutant mitigation strategy and stream credit allocation.

Contaminant transport modeling is being used more often at petroleum hydrocarbon contaminated sites in an attempt to aid engineers in evaluating the feasibility of natural attenuation as a remediation alternative in groundwater systems. In this research, a three-dimensional sequential electron acceptor computer model, SEAM3D, developed by Waddill and Widdowson (1997) was used to simulate contaminant transport at a leaking underground storage tank site in Beaufort, South Carolina. Gasoline containing benzene, toluene, ethylbenzene, and xylene (BTEX) as well as methyl tertiary butyl ether (MTBE) leaked into the subsurface at the site late in 1990, and monitoring of the water table elevations and contaminant concentrations began in 1993. Using the field data, the groundwater flow model MODFLOW was used to develop and calibrate a flow model for the Laurel Bay site using GMS (Groundwater Modeling System) v2.1. MODFLOW was coupled with the SEAM3D contaminant transport model, and the available concentration levels were used to calibrate, verify, and validate the site model. The results indicated that SEAM3D simulated complex, interconnected processes including biodegradation, and the transport of multiple hydrocarbon compounds, electron acceptors, and end products over time and space at a specific petroleum hydrocarbon contaminated site. Once the model was calibrated and verified, the model output was used to study the changes in contaminant mass distribution, contaminant mass loss, and mass loss rates for each terminal electron accepting process (TEAP) over time. It was found that the natural attenuation capacity of the aquifer was insufficient to stabilize the plume and prevent it from reaching the defined point of contact (POC). Contamination was shown to have reached the POC by 1994, just four years into the simulation. Results indicated that despite oxygen limitation within the BTEX plume, aerobic biodegradation was responsible for the greatest amount of mass loss, close to 70 %, relative to the sum of the anaerobic processes after 20 years. / Master of Science

The use of delayed drainage in land subsidence modeling greatly complicates model calibration, particularly when the thickness of the fine-grained interbeds varies throughout the modeled region. This thesis documents two separate projects (chapters) related to the use of delayed drainage in groundwater flow and subsidence modeling with parameter estimation. The overall goal of these projects was to better understand how delayed drainage affects accurate parameter estimation and how it is currently affecting the subsidence processes occurring in Las Vegas Valley. Chapter 1 describes an investigation on the value of subsidence data for groundwater model calibration considering delayed drainage. The calibration results of 13 hydraulic parameters of a synthetic conceptual model evaluated for 24 test cases indicate that (1) the inverse of the square of the observation values is a reasonable method to weight the observations, (2) spatially abundant subsidence data typically produce superior parameter estimates even with observation error under constant and cyclical pumping, (3) when subsidence data are limited and combined with drawdown data, outstanding results are obtained for constant pumping conditions. However, for cyclical pumping with observation errors, the best parameter estimates are achieved when multiple years of seasonal subsidence data are provided. The results provide useful suggestions for real-world calibration problems. Chapter 2 outlines the development of an updated flow and subsidence model for Las Vegas Valley covering the entire period of development of the basin. The new model includes a subsidence package that takes into account delayed drainage of fine-grained interbeds. Previous models used subsidence packages that assumed instantaneous equilibration of heads across all hydrogeologic units. The new model resulted in an agreement with measured water-level and improved the simulation of land subsidence. The analysis shows that the typical residual subsidence in Las Vegas Valley can be accurately simulated by incorporating delayed drainage in a long-term model. The study also indicates the need for more sophisticated modeling practices that use delayed drainage with parameter estimation processes to accurately calibrate flow and subsidence models. / Master of Science