<p> Percolation rates in Managed Aquifer Recharge (MAR) facilities, such as recharge basins and stream channels, can vary widely through both time and space. Natural variations in sediment hydraulic conductivity can create &ldquo;dead zones&rdquo; in which percolation rates are negligible. Clogging is a constant problem, leading to decays in facility percolation rates. Measuring percolation rate variations is important for management, maintenance, and remediation of surface MAR facilities. </p><p> We have used Fiber Optic Distributed Temperature Sensing (FODTS) to monitor percolation in a long narrow river channel separated from an active river by a levee. The alluvial sediment in the river channel varies widely in texture and water balance is difficult to monitor independently. The off-river channel was monitored by installing a fiber optic cable in the subsurface and measuring the propagation rate of the diurnal temperature oscillations carried downward with infiltrating water. In this way, heat was used as a tracer of percolation rates along the section defined by the 1800 meters of buried cable. We were able to confirm the FODTS measurements of percolation in the Off River Channel and demonstrate its wide applicability. Results from the measurements have been used to understand both the hydraulic behavior of percolation in the facilities and to make management decisions regarding facility operations and the potential need for additional surface sediment remediation.</p><p>

<p> Multi-model analysis (MMA) considers multiple model interpretations of a system. MMA provides a more realistic assessment of uncertainty associated with model predictions because both uncertainty of individual models and uncertainty associated with different model structures are considered. Models are evaluated for the strength of evidence that they represent an unknown system using different Information Criteria (IC) equations. IC equations are designed to assess the likelihood that a model in a set of models represents the true but unknown system. IC equations do not include a component which identifies a deficient model. Therefore, inclusion of deficient models in the set of models leads to poor model-averaged results. Evaluation of models to assess whether available observation data sufficiently support the model structure is an important step in MMA. Measures for evaluating models include: 1) failure to reach proper convergence during non-linear regression; 2) unreasonable parameter estimates; 3) unreasonable confidence intervals on parameters or a coefficient of variation greater than ten for one or more parameters; 4) high correlations between parameters; 5) determinant of the correlation matrix less than 1x10^-12; 6) condition number of the Jacobian matrix greater than 2000; and 7) unreasonable confidence intervals on predictions.</p><p> Experiments presented herein are designed to evaluate how components of AIC, AICc, BIC, and KIC rank models and assign model probabilities, and to demonstrate how removing deficient models improves MMA results. Synthetic models are used to represent true but unknown systems in contrast to experimental models that are created to simulate a simplified version of the unknown system based on observation data taken from the synthetic models. AIC, AICc, BIC, and KIC generally assign high probability to deficient models. AICc generally assigns high probability to deficient models if 1) there are many observation data or 2) there are few observation data and the model fits the data well. KIC generally assigns high probability to deficient models because these models have low Fisher Information. AIC and BIC are influenced by the goodness-of-fit and are more likely to assign high probability to more complex models because these models are generally over-fitted. Removing deficient models results in improved MMA results using AIC, AICc, BIC and KIC.</p><p>

<p> The Niles Cone Basin (NCB) within Alameda County, California, contains portions of the basin under perennial and ephemeral artesian groundwater conditions. This study used 349 wells installed throughout the basin&rsquo;s four-aquifer system to delineate the spatial distribution of the 86 wells that have gone artesian between 1995 and 2015. Artesian wells within all four aquifers occur at elevations below 5.2 meters above sea level (MASL) but predominantly below 3.0 MASL. Even at lower elevations, artesian conditions do not occur in regions of major pumping owing to significant drawdown. Within topographically-low regions, wells may not be artesian where well-heads are located at higher elevations, such as on a levee or other elevated landforms. This can be observed throughout the Newark and Centerville aquifers where artesian wells are located near non-artesian wells during the same monitoring event. Precipitation influences artesian conditions with artesian events correlated with increases in precipitation generally during, but not limited to, the early spring months. The water levels of the shallow Newark Aquifer were found to respond independently from the three deeper aquifers. Those deeper aquifers were observed to be in hydraulic connection with one another, displaying synchronous water level changes with time across the basin.</p><p>

<p> Glacier melt affects the geochemical composition of rivers; however, quantifying the glacier contribution to subarctic watershed-scale runoff has attracted limited attention. To estimate glacier contribution, we conducted a 6-year geochemical hydrograph separation study in a geologically heterogeneous glacierized watershed in Interior Alaska. Water samples were collected daily from Jarvis Creek during late April through September. Source waters were collected synoptically each year from rain, snow, baseflow (winter discharge), and the glacier terminus discharge. All samples were analyzed for stable water isotopes and dissolved ion concentrations. Stream surface water samples have large seasonal and inter-annual geochemical variation, however, source waters show distinct chemical signatures allowing the application of a geochemical hydrograph separation model to quantify relative source contribution to lowland streamflow. Considerable inter-annual differences within source water signatures emphasize the importance in informing the model with source waters sampled for each season. We estimated a seasonal average of 35% (20 to 44%) glacier terminus discharge contribution with a daily range of 2 (May) to 80% (September). If glacier contribution was to cease completely, stream discharge would be reduced by 48% and 22% in low and high rainfall summers, respectively. Combined with the documented shrinkage of glaciers, our findings emphasizes the need for further research on glacial wastage effect on subarctic watersheds.</p><p>

Aquifers hosted in fractured crystalline rocks are generally characterized by low porosity and strongly heterogeneous and anisotropic flow paths, with flow and transport dominated by discrete fracture sets. In general, zones of high hydraulic conductivity correlate with zones of high fracture intensity and fracture connectivity. Fractured rock hydraulic conductivity, however, is not only a function of spatial fracture distributions, but also displays dynamic variability due to changes in fracture aperture with changes in effective stress, such as those due to groundwater pumping and seasonal variations in water level. Some studies suggest hydromechanical coupling plays a minimal role in hydraulic conductivity alteration at shallow depths, whereas other studies attribute hydraulic conductivity alteration directly to hydromechanical coupling, thus raising a fundamental science question: what is the role of hydromechanical coupling in shallow fractured rock aquifers? This study investigates the role of hydromechanical coupling in shallow fractured rock aquifers from 3 perspectives. First, a sensitivity study presents the results of analytical and numerical modeling to determine what key hydromechanical parameters are important in the shallow crust under realistic stress states. Results suggest that hydraulic conductivity alteration is dominated by fracture normal closure, with shear dilation playing a minimal role, and that shallow dipping fractures are likely to have the highest hydraulic conductivity relative to steeply dipping fractures due to compressional deviatoric stress states. These results suggest that depth-dependent hydraulic conductivity trends observed in nature may be due in part to hydromechanical phenomena, and thus fractured rock characterization should include hydromechanical characterization. Study two presents the results of an aquifer-scale fractured rock characterization at Gates Pond, Berlin, MA in which hydromechanical variables are constrained in the field and coupled with structural and hydraulic characterization, including stress-tests, long-term water level monitoring, isotope analysis and earth tide analysis, to provide insight into the mechanical properties of the aquifer. Results of this study reveal an interesting field setting in which the mechanical properties of the aquifer are homogeneous throughout the study area, but the aquifer is compartmentalized by foliation parallel fractures that restrict hydraulic connection between wells that are placed perpendicular to foliation. Gates Pond also displays a hydraulic conductivity trend in which Foliation Parallel Fractures (FPF) have a decreasing transmissivity with increasing depth, and tectonic fractures display a decrease in transmissivity with increasing dip. Such observations suggest a conceptual model in which FPFs dominate flow in the shallow subsurface, with transmissivity decreasing with depth, where Tectonic fracture become the dominate flowing set. These results are consistent with the results of analytical and numerical modeling predictions from Chapter 1. Lastly, the third study presents results of a regional scale correlation of critically stressed fractures and fracture transmissivity in the Nashoba terrane, eastern Massachusetts. Whereas Chapter 1 suggests that fracture transmissivity is strongly modified by fracture normal closure, which is supported by field observations in Chapter 2, many workers suggest that flowing fractures are those that are critically stressed, and are thus strongly modified by shear dilation. This study addresses the role of shear dilation by identifying critically stressed fractures at a regional scale and correlating resolved stresses on transmissive fractures to fracture transmissivity. Fracture characteristics, transmissivity and borehole breakouts are characterized for 17 wells from throughout the Nashoba terrane. Critically stressed fractures are identified using inferred stress states, and correlation of critically stressed fractures to fracture transmissivity is investigated. Results suggest that transmissivity is weakly correlated to the ratio of shear to normal stress, and that ratio is strongly correlated to fracture dip. A conceptual model is proposed in which shallow dipping fractures are more likely to be critically stressed, such as FPFs in the shallow subsurface; however, high transmissivity fractures need not be critically stressed. Thus, it is concluded from observations in this dissertation that fractures in the shallow crust are most sensitive to fracture normal closure, although shear dilation may enhance transmissivity. The complex interaction between normal closure and shear dilation results in shallow dipping fractures being the most transmissive in the shallow subsurface, with tectonic fractures becoming more important with increasing depth. Each of the 3 studies presents a unique contribution to the study of hydromechanical coupling in fractured rock aquifers, with each study supporting the hypothesis that hydromechanical coupling may alter hydraulic conductivity of fractures in the shallow subsurface, contributing to observed depth-dependent hydraulic conductivity trends, variable hydraulic conductivity as a function of fracture dip, and dynamic permeability. Results of these studies show that hydromechanical coupling affects hydraulic conductivity of fractures in the shallow crust, and should therefore be incorporated into fractured rock aquifer characterization in conjunction with standard structural and hydrogeologic characterization.

No description available.

Pathogen contamination is a primary source of surface water quality impairment in the United States and modeling tools to predict concentrations of indicator organisms have utility for those involved in watershed management and ecosystem restoration, as well as public health and source and recreational water protection. This research evaluated the use of artificial neural networks (ANNs) for simulating concentrations of fecal indicator organisms in the surface waters of the Gates Brook and lower Charles River watersheds in Massachusetts. ANN model performance was assessed in terms of both the ability of the ANN models to accurately match observed indicator organism concentrations, as well as the ability of the models to correctly predict when a relevant water quality standard is violated due to exceedance of an indicator organism concentration. In addition to the fundamental question of ANN model performance, several other issues related to the development and implementation of the ANN models were explored, including the effects of different methods Of input selection and logarithmic transformation of data, the temporal transferability of the ANN models, and the potential for use of ANN models in ungaged watersheds. A comparison of ANN and ordinary least squares (OLS) regression models for fecal coliform prediction was also performed. In both watersheds, for all combinations of input parameters considered, values of average absolute percent error (AAPE) and root mean square error (RMSE) are high. When the model performance assessment is based on the ability to identify violations of a water quality standard, the impression of model performance is quite different. The best performing ANN models in each watershed are able to predict approximately 60% to 90% of the violations of the 200 CFU/100 mL standard, which is the fecal coliform standard for primary contact recreation in Massachusetts. The ANN models developed for Gates Brook show performance that is comparable to and, in some situations, slightly better than other ANN and regression models developed for indicator organisms and pathogens.

This dissertation is a composite of several studies. First, three of the most common fracture sampling techniques are tested against each other to evaluate the effectiveness of each method to adequately capture the properties of natural discrete fracture networks (DFNs). Numerical simulation is used to evaluate the single scanline, selection and multiple scanline methods in layered rocks. Using statistics from each of the techniques, DFNs are stochastically generated and compared to another network that represents the natural DFN. This model was built with the exact locations, sizes and orientations of fractures as the natural network. Porosity and permeability results reveal that the most effective method to use is the selection method because this method is consistent and performs as well as the other methods but with less expenditure of time and energy. Second, the influence of lithology and rock fabric on fracture attribute distribution in crystalline rock is assessed. Trace lengths, spacings and orientations of joints and foliation-parallel fractures (FPFs) are used to determine the potential influence of fracture type and distribution on the groundwater flow system. Results show that although both joints and FPFs are common, major orientations and spacings are different for both fracture types. Because FPFs possess identical trace lengths but narrower spacing than joints, numerical modeling experiments indicate that they play an important role in controlling the groundwater flow regime by enhancing the transmissive properties of rocks. Third, conceptual models of DFNs that have unique hydraulic character that are based on fracture configurations and properties are developed with the aid of numerical simulations. Sets of persistently parallel fractures are stochastically generated to assess the effects of fracture size and distribution, intensity, number of sets and intersection angle on the hydraulic properties of DFNs. Arbitrarily chosen class intervals of the ratio of DFN permeability to that of a single fracture are used to delineate DFNs with similar hydraulic character. Numerous graphs are created for use in the field to determine and delineate DFNs with distinct hydraulic character.

Hydrological and geochemical data was collected from a small area within the 7.32-km$\sp2$ Cadwell Creek drainage basin in west-central Massachusetts for a period of one year. The hydrologic monitoring network included a tipping-bucket rain gauge, soil-moisture and soil-temperature probes, and continuous water-level recorders to measure changing groundwater levels and stream stage. The geochemical sampling network consisted of precipitation and throughfall collectors, soil water collectors, shallow and deep groundwater monitoring wells, and stream sampling locations. Water samples were collected on a weekly or bi-weekly basis for the entire study period. Both the hydrologic and geochemical information collected during the study period indicated that the majority of water reaching the stream, both during periods of high water level and storm flow and during baseflow periods in the summer months, was derived primarily from the top meter or two of the shallow water-table aquifer. Deeper groundwater exhibited a substantially different chemistry from that in the top 1-2 meters of the aquifer. The geochemical evidence clearly indicated that deep groundwater never entered the stream, even during baseflow periods. Instead, the stream in the vicinity of the study site became dry in the summer. Comparison of groundwater and stream chemistry during periods of high water level clearly indicated that the water in the stream was derived almost solely from the shallow groundwater, with little or no contribution from more dilute precipitation and soil water. Silica concentration, as well as alkalinity and pH values, proved to be an important indicator of the origin of water entering the stream. Specific factors affecting the degree to which acidic precipitation is neutralized before entering surface water within this watershed are primarily hydrologic in nature. The amount of time that precipitation water is in contact with the geologic materials prior to entering the stream appears to determine the degree to which it is neutralized. In this case, the depth to the top of the water table is a controlling factor. Silicate weathering dominated cation exchange as a neutralization mechanism in this watershed, at least in the vicinity of the instrumented site.

No description available.