Stratigraphic and structural analyses combined with electrical resistivity soundings and drilling were used to investigate factors causing the development of sinkholes at a site in Northampton County, Northeastern Pennsylvania. By compiling and contrasting data obtained from these investigative techniques, it was possible to hypothesize the manner in which local carbonate geology determines the development of sinkholes in the area. Bedrock at the site is comprised mostly of the Epler Formation, which contains both dolomite and limestone, and is covered by glacial till which varies in thickness between meters and tens of meters. It seems, however, that sinkholes in the study area occur only in the limestone units. In areas underlain primarily by dolomite no sinkholes were detected. Sinkhole development appears to be related to the greater solubility of limestone relative to dolomite. This solubility difference produces pinnacled bedrock topography in areas underlain by limestone, in contrast to the low relief subsurface topography in areas underlain by dolomite. Electrical resistivity measurements, while not capable of distinguishing limestone from dolomite directly, can distinguish the two types of bedrock topography. Although resistivity was not able to conclusively detect clay and water filled voids directly, it was able to detect air-filled features within the bedrock. I conclude that resistivity can be used to ascertain whether the subsurface geology is characteristic of bedrock prone to the development of voids, and to decide where to drill when searching for direct evidence of voids, but only with a thorough understanding of the local geology. / Earth and Environmental Science / Accompanied by one .pdf file: 1) Jenkins-Supplemental-1999.pdf

This study created one and two-dimensional fracture flow and transport models that were compared to the transport properties of a known fracture system. In a cooperative study with Dr. Gerilynn Moline at Oak Ridge National Laboratory, Oak Ridge Tennessee, her lab data and my modeling were compared. Dr. Moline excavated columns of saprolite from Oak Ridge, and packed them in PVC pipe. The saprolite columns were flushed with tracers to study the effects of fracture flow and solute transport. Measurements taken of the column by Dr. Moline as she dissected it indicate that the fracture system was constructed of mostly small fractures (aperture = 0.02 cm), spaced close together (spacing = 0.75-1.0 cm). Tracer concentration information was collected, and Dr. Moline plotted breakthrough curves from the observed data. At Temple University, groundwater models of the saprolite columns were created using fracture flow transport programs (CRAFLUSH and FRAC3DVS). Both one and two-dimensional models were generated using data collected from Oak Ridge, and from sensitivity analyses performed at Temple. Breakthrough curves created from the modeled data were then compared to Dr. Moline's observed breakthrough curves. Results from one and two-dimensional models demonstrate that the observed fracture network does not recreate the breakthrough curves and is not representative of the dominant fracture network for the system. Modeled breakthrough curves demonstrate that the best fitting parameters create a fracture pattern made of a few large fractures (aperture = 0.16 cm) spaced farther apart (spacing = 1.5-2.0 cm), regardless of fracture lengths. The modeled fracture aperture and fracture spacing indicates that solutes flow preferentially through large fractures rather than through small disconnected fractures. This modeling method gives evidence that approximately 20% to 25% of the fractures within a system are influential in solute transport. Thus the fracture flow models calculate effective parameters that can be used to characterize the dominant fractures within a fractured porous medium. / Earth and Environmental Science / Accompanied by one .pdf file: 1) Koch-Supplemental-1999.pdf

Allocyclic analysis of the stratigraphic interval encompassing the Tonoloway­-Keyser formation boundary in Central Pennsylvania reveals a complex stratigraphic relationship produced by the superposition of orbitally forced hierarchic eustatic sea-level fluctuations on a differentially subsiding basin. To the north (Tyrone, Pennsylvania) the formation boundary is an unconformable third-order sequence boundary at which open shelf Keyser carbonate facies lie directly on peritidal carbonates of the Tonoloway Formation. Below the unconformity differential subsidence and/or uplift resulted in various amounts of vacuity throughout the study area. To the south (Everett, Pennsylvania) additional Keyser cycles occur above the unconformity, indicating decreasing hiatus in regions of greater subsidence following exposure and erosion. Beneath the unconformity, peritidal Tonoloway facies to the north grade laterally into subtidal Keyser facies to the south within individual sixth- and fifth-order cycles as a result of higher subsidence rates toward the basin axis. A correlative hierarchic allocyclic fabric of fourth- and fifth-order sequences permits a chronological analysis of these lateral facies changes and of the unconformity. Immediately before accumulation of the study interval, peritidal (Tonoloway) facies existed throughout the study area. Above this datum, over a period of 400,000 years, differential subsidence generally produced Keyser facies to the south and east, a pattern interrupted every 100,000 years by fifth-order lowstands in sea level that produced Tonoloway facies. At localities closest to the basin margin Tonoloway facies persisted throughout the study interval. If the interpretation of the hierarchic cyclic fabric is correct, at least ten fifth-order sequences ( 1 million years) are lost by erosion at some localities beneath the unconformity (vacuity) and at least five fifth-order sequences (500,000 years) are missing by non-deposition above the unconformity (hiatus). Marked differences in vacuity at Tyrone and the localities immediately to the south and east suggest Upper Silurian movement along a structural lineament. The demonstration of a basinwide correlative allocyclic framework allows quantification of the roles of tectonics and eustasy in the accumulation of stratigraphic sequences. / Earth and Environmental Science / Accompanied by one .pdf file: 1) Sinclair-Supplemental-1999.pdf

Allocyclic analysis of a 25-meter interval of Upper Silurian peritidal siliciclastic and carbonate facies of Maryland and central Pennsylvania reveals a four-tiered hierarchic cyclic fabric correlative over 125 kilometers. The structure of this correlative framework in the Williamsport Member of the Bloomsburg Formation and the Lower Wills Creek Formation is consistent with predictions of the orbital forcing model, suggesting that precession was the fundamental cycle-producing mechanism and that eccentricity modulated precession at three scales. Fundamental meter-scale cyclicity of the study interval was produced by rapid sea-level changes caused by the 20 thousand­-year precessional cycle of the Earth's orbit. These sixth-order cycles are distinguished by disjunct facies changes at non-depositional surfaces produced by rapid sea-level rises; cycles typically contain sandstone and/or carbonate at the base, and shale at the top. Sixth-order cycles are bundled by the short-eccentricity (100 thousand-year) cycle to produce asymmetric fifth-order sequences, averaging 3 meters in thickness. Eccentricity modulation also produced fourth-order ( 400 thousand-year) and third-order (2 million-­year) sequences consisting of bundles of fifth-order and fourth-order sequences, respectively. The study interval consists of two, 400 thousand-year, fourth-order sequences deposited during a period of relatively constant subsidence in shallow marine and non­marine environments. Lateral facies change within correlative cycles indicates that the southern localities were consistently more marine than the central Pennsylvania locality during the 800 thousand-years of depositional history. While subsidence and paleogeographic position contributed to the thickness of the interval and the nature of the facies, orbitally forced eustatic sea-level fluctuations were responsible for the hierarchic allocyclic framework. / Earth and Environmental Science / Accompanied by one .pdf file: 1) Zeliznak-Supplemental-1999.pdf

The lower part of the McKenzie Member of the Mifflintown Formation (Latest Wenlockian) in Pennsylvania and Maryland occurs as a hierarchically stacked set of allocycles whose causal mechanism is orbital forcing. At five localities in Maryland and Pennsylvania the lower McKenzie Limestone is divisible into two fourth-order (400 ky duration) sequences, which in tum may be divided into four fifth-order (100 ky duration) cycles. The predominant cyclic pattern, averaging 8.2 feet in thickness, is a highly asymmetric (shallowing-upward), 100 ky, fifth-order cycle. The lower parts of 100 ky cycles consist of thin (2 - 6 inches) beds of micrite, fine to coarse calcarenite and intraclastic conglomerate separated by variably thick (0.5 < 2.5 feet) intervals of gray to black shale. The upper portions of fifth-order cycles are dominated by shale, containing fewer and thinner limestone beds. Within these sequences, smaller-scale, sixth-order cycles are typically initiated by coarse calcarenite or micrite beds. Fifth-order cycles can be grouped into two fourth-order sequences averaging 33 feet thick. These larger fourth-order sequences are also asymmetric, consisting of limestone-dominated fifth-order cycles in their lower portions and shale-dominated fifth-order cycles in their upper portions. Correlation of these cyclic patterns over a distance of 160 miles indicates a eustatic causal mechanism. Specifically the predominant 6.5 - 10 feet thick fifth-order cycles are interpreted as a response to the short-term (100 ky) eccentricity mechanism and the larger composite sequences are thought to represent the long term (400 ky) eccentricity cycle. The fundamental building block of these sequences is the sixth-order cycle produced by the 20 ky precessional cycle. The hierarchic structure of these correlative patterns suggests a hierarchic mechanism, specifically Milankovitch orbital forcing. / Earth and Environmental Science / Accompanied by one .pdf file: 1) Clark-Supplemental-2000.pdf

At a contaminated site near Oyster Virginia, spontaneous potential (SP) measurements correlate closely with dissolved oxygen (DO) data from wells, indicating that a contaminant plume can be identified by this geophysical method. The field site is a 200 m X 190 m grassy field underlain by electrically homogeneous, Pleistocene, semi­consolidated sands bearing oxidized groundwater. At this site, DO measurements from 28 wells revealed a zone of reduction created by the biodegradation of buried tomato waste leaching from trenches. SP was measured using two sampling techniques: along a grid of 16 transects at 3 m intervals; and at 5 m intervals along 7 selectively oriented transects between significant wells. Spatial distribution of SP values compared favorably with the patterns formed by the DO data. This correlation was confirmed by the non-parametric Spearman rank correlation analysis, along two transects perpendicular to the redox boundary, which revealed a strong negative correlation between DO and SP at the 99% confidence level. Other potential sources of SP were ruled out by the use of resistivity, magnetometry, and ground penetrating radar, thus isolating redox potentials as the sole cause of SP. Therefore, SP can be used to map zones of reduction present in oxidized subsurfaces when all other sources of SP can be eliminated. / Earth and Environmental Science

At a site in Northampton, County, PA (Metzgar Field) of active sinkhole development (about one per month), I compared both 2D and 3D electrical resistivity for mapping the dimensions and orientations of voids within an terrain of highly variable bedrock depth using a multielectrode electrical resistivity system with a limited number of electrodes. My intention was to determine which method yields the best understanding of the subsurface and can be used to predict the potential development of sinkholes: numerous parallel 2D soundings or a single widely spaced 3D sounding. I targeted a shallow ( <30 ft deep) air-filled cavity within the carbonate bedrock of the Epler Fm. To further investigate the subsurface, I applied other methods, such as: electromagnetic induction, microgravity, drilling and the SearchCAM 2000. From the comparison, I conclude that, even though multiple 2D tests would take longer to collect data, this is the better approach to determining depth to bedrock and the possibility of cavities. This is not to say that the 2D approach is flawless. A drawback with the 2D resistivity approach is that 2D geology is assumed. This is the primary reason why the 3D approach was developed. A cavity can have an affect on a 2D test as current travels around it, even if it is offset several meters. This assumption clouds the accuracy of the 2D profile by producing anomalies in the results that may or may not be located there in the subsurface. 3D resistivity testing overcomes this assumption, but, it only enables a 7x8 electrode grid, with 56 electrodes, decreasing the electrode density along a given profile from 28 used in the 2D survey. Sounding to the depth of the cavity required an increased electrode separation from 2 m (2D test) to 10 m (3D test), resulting in an oversimplification of the subsurface, revealing the ineffectiveness of current 3D systems. / Earth and Environmental Science / Accompanied by one .pdf file: 1) Maule-Supplemental-2000.pdf

This study used various techniques to delineate groundwater flow types and the potential for fast groundwater flow in a karst system. By characterizing flow types the delineation of a wellhead protection source area was possible. Fluorescent dye tracer tests, water chemistry analyses, groundwater modeling, groundwater temperature monitoring, and geophysics were used to characterize factors which could affect groundwater flow (diffuse versus conduit flow) in a karst aquifer near Lititz, Lancaster County, Pennsylvania. Sinkholes, springs, and a well were selected for tracer test injections, water chemistry sampling, and temperature monitoring. Geophysical surveys were used to identify potential pathways (conduits) between tracer injection and sampling points at a selected site within the study area but no large conduits were found. Groundwater modeling, using the USGS software Modflow, examined how various types of karst conduit networks can lead to different sizes of source areas for wellhead protection. Two sites were studied in detail within the study area: Lititz Spring and the Buch well. Lititz Spring exhibited diffuse flow characteristics with relatively constant_ seasonal ion concentrations, diffuse tracer results, and relatively constant water levels following storm events. The Buch well site exhibited some conduit characteristics such as seasonal and time dependent characteristics with high variation in seasonal ion concentrations, high concentration of dye recovered, and rapid water level rise following storm events. The contrast in flow types at the two sites is seen in the coefficient of variation in hardness, which was 2 to 3 times higher in the Buch well than at Lititz spring. Groundwater residence times were calculated to be approximately 2 days and 9 days for the Buch well and Lititz spring respectively. Maximum dye concentrations found in tracer tests were <l µg/L for Lititz spring and estimated at >100 µg/L for the Buch well. Water levels increased by greater than 1 meter following storm events in the Buch well while water level variation at Lititz Spring was negligible (<0.25 m). The overall lack of springs and surface water features combined with field data support evidence of only localized conduit flow passages within the watershed. Groundwater modeling showed that a well-connected and expansive conduit system creates a larger wellhead source protection area as compared with only a localized karst conduit system. Tracer tests alone may not be the best way to delineate groundwater protection source areas in karst systems with a relatively localized karst conduit system. This study shows that characterization in karst systems needs to be site specific and that continuous monitoring of flow and geochemistry characteristics can be used in addition to tracer tests in delineating source protection areas. / Earth and Environmental Science

I studied the amount and type of sediment transport in a karst system as a function of storm events and seasonality. My research area is a karst groundwater system in the vicinity of the Bushkill Creek, north of the City of Easton, Northampton County, Pennsylvania. The karst system developed within the Ordovician Epler Limestone and Rickenbach Dolomite. Groundwater samples were collected from two karst springs identified along the eastern bank of the Bushkill, and a well installed in the Epler Formation, approximately 2,000 feet east of the Bushkill in a local recharge zone. Samples were also collected from a non-karst spring in the red shale of the Triassic Stockton Formation approximately 40 miles south of Easton in the Town of Buckingham, Bucks County, Pennsylvania. Biweekly samples were analyzed for particle size and water chemistry and selected samples for clay mineralogy. In addition, water level was measured in the monitoring well and Spring No. 1. An increased amount of sediment was transported in the monitoring well (approximately 200 to 800 mg/L) during the autumn and early fall compared to a low of 0 mg/L during the spring and summer. The spring discharged a low amount of sediment (a maximum of 9 mg/L) during the autumn of 2000 and three times the amount during the spring and summer. One storm monitored in March, 2000, monitored at the spring only, showed that the sediment load reached the maximum discharge of sediment (13 mg/L) subsequent to the storm peak. Correlation between sediment yield and water level in Spring No. I and the monitoring well.show that the overall sediment load increased during periods of low precipitation. The cumulative effect of karst systems flushed during the spring and summer may be to decrease sediment loading to discharge points, while water chemistry varies with individual precipitation events. The aqueous chemistry of the system shows dilution during wet periods and enriching of ions during dry periods. Storm chemistry shows that both dilution and flushing does occur in karst systems because ion concentration decreased over the course of a storm whereas the sediment yield increased. Sediment transport in karst aquifers varies because of how precipitation infiltrates the complex diffuse-­conduit systems of karst. / Earth and Environmental Science

The Brian Maebius Site in Badlands National Park, S.D. is located in the lower/middle Scenic Member of the Oligocene Brule Formation of the White River Group. The Site is a collection of fluvially reworked volcaniclastic and epiclastic sediments of mudstones, channel sandstones, siltstones, and vertebrate fossils. Macroscopic, microscopic, and geochemical data that I collected at Brian Maebius suggests that paleosols are prevalent and consist predominantly of weakly developed Haplustalfs, Haplustepts, and one pervasive, strongly developed Paleudalf, which is informally referred to as the Hay Butte Marker Bed. Data also indicates that these paleosols represent sub-humid to semi-arid and semi-forested to savanna paleoenvironments within a meandering to anastomosing floodplain setting. Taphonomic data indicates that the Brian Maebius Site contains weakly to moderately weathered, disarticulated, fluvially abraded, carnivore processed, and mostly randomly oriented vertebrate fossils that are adjacent and laterally traceable to the Haplustalfs and Haplustepts. These taphonomic indicators suggest that attritional accumulation and/or predators proximal to an oxbow setting and active stream channel were responsible for bone accumulation, and that periodic flooding events transported and buried faunal remains. This research is part of a program sponsored by the National Park Service to document and protect fossil resources within Badlands National Park. Detailed analyses of sedimentology, paleopedology, stratigraphic position, and geographic distribution of fossil sites will help to interpret the effect of basin dynamics on bone bed genesis, generate models to predict new sites, and help protect bones from poaching. / Earth and Environmental Science / Accompanied by one .pdf file: 1) Factor-Supplemental-2002.pdf