TUScholarShare
Communities in TUScholarShare
Select a community to browse its collections.
Recent Submissions
Item Open Access Ground-water model of spray irrigation in New Garden Township, Chester County Pennsylvania(Temple University. Libraries, 1999)I used numerical modeling to simulate the flow of ground water at a spray irrigation site. The USGS code MODFLOW was used to create two and three-dimensional steady-state simulations, and three-dimensional transient simulations. The purpose of the ground-water model was to investigate the complexity in the parameters used in the model. The model parameters included horizontal hydraulic conductivity, vertical hydraulic conductivity (represented by the MODFLOW variable VCONT), irrigation rate, precipitation, horizontal anisotropy (the ratio of and Ky), and bedrock topography. Site-specific data was not readily available, so there was uncertainty in parameter values. Sensitivity analysis was conducted to evaluate the complexity needed in the parameters to accurately model the hydrologic system at the spray irrigation site. Two methods of sensitivity analysis were used: systematic variation of two parameters (while holding others fixed), and Latin hypercube sampling. Both steady-state and transient simulations were performed. Transient modeling showed the same trends as steady-state modeling with the systematic variation of hydraulic conductivity and recharge. The sensitivity analysis showed that the hydraulic conductivity was an important parameter in predicting the effects of spray irrigation, where the model was most sensitive to values of hydraulic conductivity less than 25 m/day. This sensitivity to hydraulic conductivity has implications for data collection in that if the hydraulic conductivity of the top layer is 25 m/day or less, then more detail in hydraulic conductivity distribution and mean values are necessary to understand the effects of spray irrigation on the hydrologic system of the area. If the hydraulic conductivity values are found to be greater than 25 m/day, then detail is not as necessary. The three-dimensional modeling was beneficial in understanding the role that the soil plays in stream discharge from spray irrigation. Both the recharge to the regional ground-water system and the stream discharge are important in evaluating spray irrigation. Performing sensitivity analysis before site investigation can aid in the collection of field data, and to help make predictions about the effectiveness of spray irrigation at a particular site.Item Open Access Using groundwater modeling to understand the factors controlling lake seepage patterns(Temple University. Libraries, 2008)Understanding groundwater-lake interactions requires detailed knowledge of seepage patterns. Simple numerical models predict an exponential decrease of seepage with increasing distance offshore, but field measurements have shown that many lakes deviate from this pattern. To examine how hydrogeologic factors influence lake seepage patterns, two-dimensional steady state numerical models were constructed using the USGS modeling code MODFLOW. Sensitivity was tested to recharge, groundwater basin size and thickness, and lake morphology, as well as heterogeneities in lake sediment thickness and hydraulic conductivity, bedrock fracture zoning, and the effect of combining factors. The models were hypothetical. but many of the hydrogeologic parameters were based on those seen at Mirror Lake, a small glacial lake in New Hampshire. The models showed that increasing the thickness of lake bed sediments or decreasing their conductivity, increasing aquifer anisotropy, and decreasing the penetration of a lake into the groundwater system resulted in seepage rates that were more evenly distributed across a lakebed rather than focused near shore. Changing overall lake slope without increasing the penetration of the lake into the aquifer system produced little noticeable change in seepage patterns. Localized geologic heterogeneities, including a pinch-out in lake sediment and the presence of a high-K fracture zone, as well as a change in lake slope produced localized peaks in seepage. The peak produced by a pinch-out in lake sediment increased in magnitude as the degree of thinning increased, with seepage rates that locally were magnified by a factor of 150 where lake sediment was absent. A high-K fracture zone produced a less significant peak whose size was controlled by several factors, including the overall length of the fracture, its connectivity to the lake, its orientation, and the presence or lack of low-K lake sediment. A change in lake slope produced a smaller peak compared with the pinch-out and fracture zone cases except when modeled using highly anisotropic conditions in which seepage rates were magnified by a factor of 20. Heterogeneities modeled closer to shore are more likely to contribute a significant amount of seepage to a water budget than those further from shore. One important factor controlling the degree of change in seepage was the contrast between lake bed sediments and the bedrock aquifer. Thus, the conductance of lake bed sediment and anisotropy should be measured to predict whether heterogeneities or breaks in lake slope will produce detectable localized changes in seepage rates. This modeling can help guide field studies by providing an understanding of the geologic conditions that control the rate of exponential decline of seepage away from shore.Item Open Access Groundwater modeling with GIS methods for the Mullica River Basin, New Jersey(Temple University. Libraries, 2006)I constructed a 3-D numerical model for the Mullica River basin, New Jersey. The Geographic Information System (GIS) method was applied in groundwater modeling to process land features into model format. Model parameters were input into Groundwater Modeling System (GMS) by using GIS features such as land use for recharge zones, and stream locations for groundwater discharge boundaries. I selected MODFLOW 96 as the modeling code. The Mullica River basin is a watershed with high vulnerability to stresses, which scored the highest level of 6 on the Index of Watershed Indicators (IWI) by EPA. The Mullica River basin is threatened by aerial and point sources of contamination. Simulation results of the groundwater model were used in Watershed Modeling System (WMS) to simulate the differences between the surface and groundwater basins, which are the boundaries of the contaminant migration. The groundwater model was calibrated to 140 observed water level and 3 continuous gaging stations. The calibrated model had 4.99 ft root mean square residual, with matched flow field. To evaluate the sensitivities of the model to the key parameters, 70 sensitivity model simulations were performed. The key parameters included hydraulic conductivities, aerial recharge, and the conductance of the riverbed. The model was sensitive to hydraulic conductivities and recharge values, but not sensitive to the river bed conductance. GIS procedures described in this study were helpful in incorporating accurate boundaries and zoning recharge based on weather, land use, and soil type. Integrating Arcview and GMS together was convenient in inputting data and post-processing model results. However there are some disadvantages in application of GIS in groundwater modeling such as lack of hydrological data on river elevations in the GIS database and inappropriate data arrangement for stream groups. The model could not be calibrated to a specific stream order, but a combination of second and third order streams were used. Watershed simulations demonstrated differences in groundwater basins and surface water basins. The surface water flow basins were slightly smaller than groundwater basins. Particle tracking showed that flow paths on the surface diverged from groundwater paths at a few locations. Development in the Pine Barrens should consider the environmental impact based on surface water boundaries as well as on groundwater boundaries.Item Open Access Characterizing Storm Response in an Urban Karst Aquifer at Valley Creek, Chester County, Pennsylvania(Temple University. Libraries, 2006)Karst aquifers respond more quickly to precipitation events than other types of aquifers. However, continuous monitoring and geochemical analysis can provide the necessary information to better understand aquifer recharge pathways and sources during precipitation events. From May 2005 to March 2006 continuous data and water samples were collected from four springs in the Valley Creek Watershed, Chester County Pennsylvania. The springs selected for this study are in commercial or residential land use in an urban setting. This area is underlain by carbonate units that trend east-northeast to west-southwest bounded on the north and south by crystalline igneous and metamorphic units. All springs were in the Elbrook Formation, which is a fractured dolomite unit of Upper Cambrian age. The goal of this study was to look at the effects of urbanization in a karst setting by monitoring conductivity, water level, and temperature at several springs with different land-use and use halides to distinguish possible source of anthropogenic contamination of the groundwater. Storm response was rapid, on the order of 1-3 hours. Recovery to baseflow conditions of both conductivity and water level was usually within 3-6 hours. This was in contrast to storm response measured in a previous study in a rural setting showing 1 -2 days response time from the start of the precipitation event to increases in water stage and to decreases in conductivity at the monitored sites. The water stage response was also on the order of 10-100 centimeters higher in the rural setting than in the urban setting. The maximum water stage increases at the urban springs were typically less than 10 centimeters greater than baseflow conditions, with the more commercialized springs showing only < 5 cm increase. The fast response and lower water stages at these urban springs were due to smaller capture areas for the springs compared to those in the rural settings. Conductivity responses were small during storms similar to water level responses. The timing of the initial conductivity responses corresponded also with the timing of the initial water level responses. In general a decrease in conductivity occurred during storm responses, this indicated that recharge moves through quickly. There were slight variations in storm responses between the different springs. Overall however, each spring showed little variation from storm to storm. The consistent and fast response indicated short, fast flow paths. Geochemical analysis of halide concentrations (chlorides and bromide) in the water samples from the commercial spring site shows contamination by road salt. High chloride concentrations (150 mg/L -280 mg/L) were measured throughout the year. The excess application of road salt during deicing periods has led to accumulation of road salt in the capture area of these springs and has become a continuous anthropogenic halide contamination source. The urban setting thus influenced spring response by limiting the capture area and reducing infiltration. The reduced infiltration led to smaller storm response and the smaller capture area led to rapid storm response. The presence of high chloride concentrations, concluded to be contributions from road salts also was further indication of shallow, rapid flow paths in this urban setting.Item Open Access Analysis of Effects of Urbanization on Spring Geochemistry and Sediment for a Karst Aquifer in Valley Creek Basin, Chester County, Pennsylvania(Temple University. Libraries, 2007)Valley Creek Basin located in eastern Chester County, southeastern Pennsylvania, is an urban karst system. Four springs within the Elbrook Formation and two springs in the Ledger Dolomite have been monitored for suspended sediment, baseflow water chemistry, and storm response. Data from previous studies at two rural sites, Nolte in Lancaster and Bushkill in Easton Pennsylvania, were used for comparison to examine the effects of urbanization on spring chemistry. The flow paths were compared using hardness coefficient of variations (CV) to differentiate between matrix and conduit dominated springs. The urban springs CV values ranged from 14.5% to 8.2%, showing a mixture between conduit and matrix flow paths. The rural sites CV values place Nolte as a matrix dominated system (5.6%) and Bushkill as distinctly conduit (17.5%). The nitrate values for the urban sites were low (< 5 mg/L) and had little variation within a 1 to 2 mg/L range. The rural sites had higher nitrate concentrations that showed seasonal variations around 10 mg/L. The difference between the urban and the rural nitrate values is most likely due to the agricultural use of fertilizers in rural areas. The calcite and dolomite average Si’s were slightly undersaturated (-0.01 to -0.5), suggesting that these systems are most likely made of solutionally enlarged fractures, with mixing between conduit and matrix flow. Most of the urban log PCO2 values averaged around -1.9, indicating that the urban springs are open systems under the influence of soil gas. The Bushkill spring’s log PCO2 averaged -2.3 indicating that it is also an open system, while Nolte spring had slightly higher log PCO2 (around -1.5) indicating a closed system. The suspended sediment concentrations for the urban springs have shown fairly constant low levels. The concentrations mostly range from 0.5 to 4.0 mg/L. The low suspended sediment values suggest that the karst system is not receiving a heavy load of urban sediments in the recharge area. It is believed that urbanization and increased impervious surfaces have lead to a reduction of infiltrating waters, which has in turn limited the amount of sediment entering the system. Furthermore, trace metal analysis of baseflow water samples and spring mouth sediment showed low concentrations. The metal concentrations in the water samples were generally less than 10 ppb and below EPA drinking water standards. The metal concentrations in the sediment were higher than those in the water samples. Even so, two-thirds of the sites had no metal concentrations over typical background values. The low concentration of metals on the sediment suggests that there is a lack of build up of metals in the karst system. Karst aquifers are usually more vulnerable to contamination than other aquifers due to reduced filtration of infiltrating waters. However, urbanization in Valley Creek Basin’s karst system has not had the effect of exposing the karst system to metal or sediment contamination, although the other geochemical indicators are typical of karst with a mixture of matrix and conduit flow paths. The lack of metal and sediment contamination appears to be related to urbanization restricting the recharge area and thus the potential flux of sediment and metals.