Prediction of Remediation of a Heterogeneous Aquifer: A Case Study

Contaminant plumes whose characteristic length is smaller than the horizontal integral scale of the hydraulic conductivity, K, are abundant in shallow, phreatic aquifers. In such cases, the aquifer can be regarded as layered, with K being only a function of the vertical coordinate. The heterogeneity of K has a critical role upon the efficiency of remediation of such sites, for example, by Pump and Treat schemes. The expected efficiency is a random variable, with uncertainty. Quantifying this uncertainty can be of great importance to decision making. In this study, we focus on a case study in the coastal aquifer of Israel and compare two different approaches for constructing realizations of K: continuous and indicator. We observe a significant difference between the constructed realizations, which results in a considerable difference in the predicted remediation efficiency and its uncertainty. Furthermore, we study the effect of conditioning the realizations by a rather limited number of K data points. We find that the conditioning results in a major reduction of the uncertainty. In addition, we compare the results of the transport model to a simplified semi‐analytical solution that is based on assuming radial flow. We find a good agreement with the three‐dimensional numerical model. This result illustrates that the simplified solution can be used for prediction of the remediation efficiency when the flow at the plume vicinity can be regarded as radial.     

DNAPL Flow Behavior in a Contaminated Aquifer: Evaluation of Field Data

A pool of dense nonaqueous phase liquid (DNAPI.) containing TCE and other chlorinated solvents has been removed from the subsurface at Hill Air Force Base, Uthah. as part of an interim remedial action. The removal of the DNAPI. pool means that future off-site migration of dissolved contaminants in the ground water is minimized, and costs for final remedial actions are reduced. A pump-and-treat system recovered more than 23.000) gallons of DNAPI. and one million gallons of contaminated ground water from the aquifer. The efficiency of this remedial action was evaluated on the basis of extensive field and laboratory data. The behavior of DNAPI. flow in the aquifer sands was characterized by collecting core samples from two borings in the DNAPL pool and measuring relative permeabilities and DMAPI. saturation. Core Hooding results show that approximately one-third of the DNAPI. originally in the pool is not recovered by water displacement, but remains as a residual saturation held in place by capillary pressure. However, subsequent Hooding with two pore volumes of surfactant solution reduced the residual DNAPI. saturation in the sand by one order of magnitude. Analytical and numerical models for the DNAPI flow behavior at the site were developed. This is the first time that such models have been developed and applied to an actual DNAPI. pumping lest conducted in the field. Because measured permeabilities and residual saturations were used lo calibrate the models. the model predictions could be used lo provide valuable insights into the controlling mechanisms for DNAPL recovery. The data collection and modeling procedures outlined in this paper can be used lo enhance the efficiency and minimize the cost 10 clean up this and other DNAPI.-contaminated sites.     

Submarine Ground Water Discharge Driven by Tidal Pumping in a Heterogeneous Aquifer

Submarine ground water discharge (SGD) is now recognized as an important water pathway between land and sea. It is difficult to quantitatively predict SGD owing to its significant spatial and temporal variability. This study focuses on quantitative estimation of SGD caused by tidally induced sea water recirculation and a terrestrial hydraulic gradient. A two-dimensional hydrogeological model was developed to simulate SGD from a coastal unconfined aquifer in the northeastern Gulf of Mexico, where previous SGD studies were performed. A density-variable numerical code, SEAWAT2000, was applied to simulate SGD. To accurately predict discharge, various influencing factors such as heterogeneity in conductivity, uncertain boundary conditions, and tidal pumping were systematically assessed. The tidally influenced sea water recirculation zone and the fresh water–salt water mixing zone under various tidal patterns, tidal ranges, and water table heights were also investigated. The model was calibrated and validated from long-term, intensive measurements at the study site. The percentage of fresh SGD relative to total SGD ranged from 4% to 50% under normal conditions. Based on simulations of two field measurements in summer and spring, respectively, the fresh water ratios were 9% and 15%, respectively. These results support the hypothesis that the SGD induced by tidally driven sea water recirculation is much larger than terrestrial fresh ground water discharge at this site. The estimates of total and fresh SGD are at the low and high ends, respectively, of the estimation ranges obtained from geochemical tracers (e.g., ²²²Rn).     

Removal of PCE DNAPL from Tight Clays Using In Situ Thermal Desorption

This paper presents a full‐scale thermal remediation of a brownfields site near San Francisco, California. In Situ Thermal Desorption (ISTD) was used for treatment of chlorinated solvents in a tight clay below the water table. The site had contaminants in concentrations indicating that a tetrachloroethene (PCE)‐rich DNAPL was present. A target volume of 5097 m³ of subsurface material to a depth of 6.2 m was treated for a period of 110 d of heating. Energy was delivered through 126 thermal conduction heater borings, and vapors were extracted from a combination of vertical and horizontal vacuum wells. Approximately 2540 kg of contaminants were recovered in the extracted vapors by the end of treatment. The PCE concentration in the clay was reduced from as high as 2700 mg/kg to an average concentration of 0.012 mg/kg within 110 d of heating (a reduction of >99.999%). Similar effectiveness was documented for TCE, cis‐1,2‐DCE, and vinyl chloride. A total of 2.2 million kWh of electric power was used to heat the site. Approximately 45% of this energy was used to heat the subsurface to the target temperature. Another 53% was necessary to boil approximately 41% of the groundwater within the treatment zone, creating approximately 600 pore volumes of steam by the end of the 110‐d heating and treatment period. Steam generation thus occurred within the clay. Partitioning of the contaminants into the steam and its removal comprised the dominant remedial mechanism. The steam migrated laterally toward the ISTD heaters, where it encountered a small dry region adjacent to each of the heaters, which served as a preferential pathway allowing the steam to migrate upward along the heaters to the more permeable vadose zone. There the steam was captured by a system of vertical and horizontal vacuum extraction wells. This vapor removal strategy facilitated effective thermal treatment of the tight clays located below the water table. Features of a robust design are extension of the heaters at least 1.2 m deeper than the treatment depth, and the installation of shallow horizontal vapor collection wells which allow for establishment of pneumatic control.     

Multiple Working Transport Hypotheses in a Heterogeneous Glacial Aquifer System

Multiple working hypotheses can be used to evaluate permissible alternative hydrogeological interpretations at sites with limited subsurface control. This approach was applied to test the viability of three conceptual aquifer system architecture models coupled with three hypothesized source locations for a 1,4-dioxane plume in a heterogeneous glacial aquifer system in Washtenaw County, Michigan. The three alternative conceptual models characterized the site hydrogeology with increasingly complex distributions of hydrostratigraphic units: (A) an effective aquifer, (B) a layered confined aquifer, and (C) a discretely heterogeneous aquifer model. Each was incorporated into an independently calibrated numerical ground water flow (MODFLOW) model. Steady-state and transient flow simulations of the alternative models were evaluated using both hydraulic flow field characteristics observed under natural conditions and the perturbed response after local remedial pumping activity began. Three plausible locations where 1,4-dioxane could have entered the aquifer system were identified using historical information at the site: (1) manufacturing waste water disposal lagoons, (2) a 60 foot (18 m) deep kettle lake, and (3) a shallow impoundment on a local stream. Advective transport modeling (MODPATH) was used to assess the consistency of the hypothesized source locations with observed contaminant migration pathways inferred from the mapped location of the plume. Evaluation of the nine combinations of hydrogeologic conceptualizations and 1,4-dioxane source locations led to elimination of four working hypotheses and discounting of two others, leading to reduced overall uncertainty and the development of new insights into the system behavior.     

Analytical Solution for Modeling Discharge into a Tunnel Drilled in a Heterogeneous Unconfined Aquifer

Predicting transient inflow rates into a tunnel is an important issue faced by hydrogeologists. Most existing analytical solutions overestimate the initial discharge due to the assumption that drilling was instantaneous over the entire tunnel length. In addition, they assume a homogeneous system. An alternative model was recently developed for tunnels intersecting heterogeneous formations, but its application was reduced to the case of confined flow to deep tunnels in weakly diffusive aquifers. In this paper, we adapt existing analytical solutions for drainage systems to the specific case of a tunnel progressively drilled in a highly diffusive heterogeneous unconfined aquifer. The case of a tunnel overlying an impervious layer is analytically solved by applying the superposition principle, while the case of a tunnel constructed some distance above an impervious layer is solved by discretizing the tunnel length into subsectors. Both models can simulate transient discharge into a tunnel drilled at various speeds through a heterogeneous unconfined aquifer, and allow the prediction of discharge rates in shallow tunnels located in highly diffusive aquifers. We successfully applied this approach to a tunnel in heterogeneous volcanic rock.     

Storage and Drainage Characteristics of a Highly Heterogeneous Karst Aquifer in Houzhai Basin

The characteristics of karst aquifers are difficult to be determined due to their heterogeneous physical properties and lack of hydrogeological information. In this case study, we applied two methods for a comparative analysis of storage and drainage characteristics in upstream, midstream, and downstream of Houzhai cave stream basin. In the first method, Minimum Smoothed Method (MSM) is used to determine the proportion of baseflow to the total flow (Baseflow Index, BFI). In the second method, a bicarbonate‐base two‐end member mixing model is used to quantify the slow flow component and fast flow component. For both methods, slow flow and quick flow are quantified at three sampling sites, which provide useful information for the analysis of storage and drainage characteristics. The results from flow separation method and hydrogeochemical analysis show a consistently increasing trend of the proportion of slow flow to total flow from the upstream to downstream which indicates that the voids of highly conductive conduits and well‐connected fissures decrease along the flow paths in the Houzhai cave stream basin in southwest China. The upstream areas have a low proportion of baseflow which indicates a high drainage capacity due to high permeable conduits and well‐connected fissures. The downstream areas, on the contrary, have a high proportion of baseflow which indicates a high storage capacity and slow infiltration due to the predominant presence of matrix and poorly‐connected fissures. These numerical methods provide alternative ways to investigate the storage and drainage characteristics of karst aquifers where direct measurement are not available.     

DNAPL accumulation in wells and DNAPL recovery from wells: Model development and application to a laboratory study

Dense nonaqueous phase liquid (DNAPL) accumulation and recovery from wells cannot be accurately modeled through typical pressure or flux boundary conditions due to gravity segregation of water and DNAPL in the wellbore, the effects of wellbore storage, and variations of wellbore inflow and outflow rates with depth, particularly in heterogeneous formations. A discrete wellbore formulation is presented for numerical modeling of DNAPL accumulation in observation wells and DNAPL removal from recovery wells. The formulation includes fluid segregation, changing water and DNAPL levels in the well and the corresponding changes in fluid storage in the wellbore. The method was added to a three-dimensional finite difference model (CompSim) for three phase (water, gas, DNAPL) flow. The model predictions are compared to three-dimensional pilot scale experiments of DNAPL (benzyl alcohol) infiltration, redistribution, recovery, and water flushing. Model predictions match experimental results well, indicating the appropriateness of the model formulation. Characterization of mixing in the extraction well is important for predicting removal of highly soluble organic compounds like benzyl alcohol. A sensitivity analysis shows that the incorporation of hysteresis is critical for accurate prediction. Among the multiphase flow and transport parameters required for modeling, results are most sensitive to soil intrinsic permeability.