Richard R. Parizek

Relationship between fracture traces and the occurrence of ground water in carbonate rocks

Abstract Fracture traces visible on aerial photographs are natural linear-drainage, soiltonal, and topographic alignments which are probably the surface manifestation of underlying zones of fracture concentration. The relationship between fracture traces and occurrence of ground water was studied in Nittany Valley of Central Pennsylvania, U.S.A., an area underlain by folded and faulted Lower Paleozoic carbonate rocks with interbedded sandstones. Several bore-hole caliper surveys were run to determine if fracture zones underlying fracture traces localized cavernous openings. Three caliper logs, obtained from wells drilled on fracture traces, showed that numerous cavernous openings were penetrated by wells extending to 350 feet in depth, while four logs obtained from wells drilled in interfracture trace areas showed few cavernous openings. Specific capacities, unadjusted for variations in well diameter, pumping period, and well losses, were determined for eleven wells drilled in dolomite and sandy dolomite and for three wells drilled in limestone. The specific capacities were divided by the total depth of saturated rock penetrated by each well to obtain the specific capacity in gallons per minute per foot of drawdown per foot of saturated rock penetrated. These values ranged from .0008 and .002 for two wells drilled in dolomite in interfracture trace zones to .02, .04, .04, .07, .08 and .09 for six wells drilled on or near a single fracture trace in dolomite, 1.52 and 1.23 for two wells drilled on a single fracture trace in limestone, .11, .12 and .17 for three wells drilled in dolomite at or near the intersection of two fracture traces and 3.27 for a single well drilled at the intersection of two fractures in limestone. These data support the concept that fracture traces reflect underlying fracture concentrations and are useful as a prospecting guide in locating zones of increased weathering, solutioning and permeability.

Soil-water sampling using pan and deep pressure-vacuum lysimeters.

Abstract A description of two soil-water sampling devices and success achieved in obtaining soil water on a routine basis at depths of 1 to 36 feet below land surface are described. In one facility, galvanized, 16 gage metal pans, 12 × 15 inches, with a copper spout, are driven into the wall of an open trench to intercept gravitational water at 1-foot intervals to a depth of 17 feet. Water samples are diverted to sample bottles attached to the inside walls of a protective housing by way of plastic tubing. Flooring is required to prevent quicking of stratified residual silt loam to sandy loam soils at the base of the trench. From 4 to 6 inches of water are irrigated adjacent to the trench on a weekly basis and a floor drain or sump pump is required to prevent flooding of the sampling trench. The deep trench lysimeter provides gravitational water samples on a routine basis when waters are artificially applied to the site but failed to yield samples during the growing season in the absence of irrigation under Pennsylvanias climate. A modified version of a commercially avaiable lysimeter constitutes the second device which can extract soil water under tension from depths in excess of 50 feet. The commercial lysimeter contains a porous ceramic cup attached to a 2-foot section of plastic pipe with a one-hole rubber stopper. In the modified version a two-hole rubber stopper is used and two copper tubes are inserted through the stopper, one for evacuating the tube, the other to force out water. This improved soil-water sampler has been named a pressure-vacuum lysimeter or “suction lysimeter.” One or more lysimeters may be installed in a 6-inch diameter drill hole. The ceramic tip is embedded in a pulverized silica deposit to insure hydraulic continuity with the soil-water reservoir. Native soil, bentonite plugs, or grout plugs are used to backfill holes to prevent channeling. A hand pump may be used to evacuate the tube and to blow out samples. When sufficient soil-moisture is available 500 to 970 mls of water may be obtained on a weekly basis. During prolonged dry periods, no samples can be collected when soil water has been depleted around the ceramic point. This may be four to eight weeks after pan lysimeters fail to yield water samples.

Numerical simulation of the Noordbergum effect resulting from groundwater pumping in a layered aquifer system

Abstract Numerical simulations were carried out to investigate the so-called Noordbergum effect (i.e. reverse water-level fluctuation) resulting from groundwater withdrawal. Two specific cases were analyzed: a three-layer aquifer system composed of two permeable aquifers separated by a semi-impermeable aquitard, and a corresponding single-layer aquifer system composed of an equivalent lumped material. In the numerical simulation of the layered aquifer system, the Noordbergum effect is observed during the early time period of pumping in the overlying aquitard and unpumped aquifer while the underlying aquifer is pumped. In contrast, the numerical simulation results of the lumped aquifer system do not show such a reverse water-level fluctuation throughout the entire domain. These results strongly suggest that the Noordbergum effect can be explained by the difference in poroelastic responses of the layered (heterogeneous) and lumped (homogeneous) aquifer systems to the hydraulic pumping stress. The Noordbergum effect is caused by two mechanisms: a faster mechanical propagation (deformation) of the pumping stress than its hydraulic propagation (drawdown) from the pumped aquifer into the adjacent aquitard and unpumped aquifer due to relatively lower hydraulic conductivity of the aquitard, and a mechanical amplification (excessive compression) in the lower part of the relatively soft aquitard. However, the pumping stress is evenly distributed throughout the entire domain of the lumped aquifer system without such mechanisms since it has homogeneous hydraulic and mechanical properties.

Three-dimensional finite element modelling for consolidation due to groundwater withdrawal in a desaturating anisotropic aquifer system

A poroelastic numerical model is presented to evaluate three-dimensional consolidation due to groundwater withdrawal from desaturating anisotropic porous media. This numerical model is developed based on the fully coupled governing equations for groundwater flow in deforming variably saturated porous media and the Galerkin finite element method. Two different cases of unsaturated aquifers are simulated for the purpose of comparison: a cross-anisotropic soil aquifer, and a corresponding isotropic soil aquifer composed of a geometrically averaged equivalent material. The numerical simulation results show that the anisotropy has a significant effect on the shapes of three-dimensional hydraulic head distribution and displacement vector fields. Such an effect of anisotropy is caused by the uneven partitioning of the hydraulic pumping stress between the vertical and horizontal directions in both groundwater flow field and solid skeleton deformation field. Copyright

The hydrologic catchment area of a chain of karst wetlands in central Pennsylvania, USA

Shallow depressions found in karst terrains may contain wetlands (karst pans) that fluctuate seasonally in response to climatic conditions. This study examined the ground-water hydrology of a chain of 17 wetlands, ranging in size from 0.06 to 0.4 hectares, located in an Appalachian karst valley in central Pennsylvania, USA. The study objective was to determine the contributing area of wetland source waters. A variety of hydraulic head, soil stratigraphy, and water chemistry data indicate that the combined contributing areas of these wetland ponds extend to a maximum distance of approximately 150 m from the ponds. The hydrologic catchment area of the ponds during January and February, 1999 (approximately 2–8 hectares) was significantly smaller than catchment area based on topography (69 hectares). The water source of the wetlands consists of direct precipitation inputs and shallow (0.5–6 m) perched ground water. The hydrologic catchment area of the ponds expands during wet periods and contracts during dry periods. The ponds have been shown to be perched above the regional water table.

An application of parametric statistical tests to well-yield data from carbonates of central Pennsylvania

Abstract Variation in productivity, yield in gallons per minute per foot of drawndown per foot of saturated thickness, of 80 water wells located in folded and faulted carbonate rocks and shales of Cambro-Ordovician age in central Pennsylvania, was related to six hydroecologic factors. Parametric and nonparametric statistical techniques were applied and results of One Factor Analysis of Variance and Student t-test are presented. Productivity values were transformed to common logarithms, and it was assumed that a log-normal model would reasonably describe the variation in productivity especially as the sample size was increased. These tests showed that the variations in number of fracture traces, rock type, dip of bedrock strata and topography were significant and variation in depth to water table was not significant in accounting for variation in well yield. Student t-test showed that anticlinal wells were significantly different from synclinal wells. However, wells in the same rock type but different structural settings were not significantly different. This shows that variations in rock type and number of fracture traces are more important than other structural variations. Both parametric and nonparametric tests gave identical results, which justifies the use of parametric tests which require normally distributed data.

Groundwater tracing with post sampling activation analysis

Abstract A tracing technique is described which uses low concentrations of a nonradioactive tracer, the bromide ion. Concentrations may be so low that they do not constitute a health or pollution problem when added to segments of the hydrologic cycle. The tracer is detected by post sampling activation analysis techniques. Water samples containing the tracer are made radioactive when irradiated with neutrons. Radiation detectors are used to analyze the radiation emitted by each sample and the unique radiation characteristics of the tracer ( 80 Br with a half life of 17.6 min) are sought. No chemical separation techniques are required. Given a “rabbit system” to transport irradiated samples to a Ge(Li) detector 3–5 samples per hour can be analyzed. The technique is safe, and tracers can be detected as low as 20 parts per billion (p.p.b.) over that of background concentrations. The bromide ion does not appear to be lost by precipitation, absorption or adsorption and is biologically stable. A spike concentration of 200 p.p.m. was used in field studies because it is lower than the limit set for drinking water. The technique was tried on sinkhole, conduit and spring systems in a carbonate terrain, in central Pennsylvania. Surface water entering a sinkhole had a background bromine concentration of 0.020 μg/ml and 0.027 μg/ml at the conduit outlet 0.6 miles distant. Although the flow rate into the sinkhole was 50 ft. 3 /min, no tracer was detected in the spring 16 h after the tracer was added. In a second study ammonium bromide was added in a conduit 500 ft. from its discharge point. Tracer first appeared at the discharge point 1 h after injection (0.11 p.p.m. vs. an average background of 0.03 p.p.m.). The maximum passed at 1 h 30 min and the spread of the spike was estimated to be over 5 h. For a peak maximum of 0.1 p.p.m., only 3 g of ammonium bromide would have been required in the second study. The inlet concentration would have been only 2.8 p.p.m. indicating the sensitivity of the method. Nine of many possible potential field applications are listed for this superior tracing technique.

Geological controls on seasonal-pool hydroperiod in a karst setting

Shallow depressions found in karst terrains may contain temporary (vernal) pools that are inundated seasonally in response to changes in meteorological conditions. The hydrogeology of 16 pools (0.06–0.4 ha) was studied in an Appalachian karst valley in central Pennsylvania, USA. The objective was to determine the effect of the geologic substrate on pool hydroperiod. Meteorological, geophysical, and hydrogeological data collected from November 1997–August 1999 and from January 2002–January 2004 suggested that hydroperiod was primarily controlled by meteorological conditions (total annual precipitation) and surficial aquifer geology. Multiple regression models were found to predict most of the spatial variability of pool hydroperiod with the following variables: thickness of the surficial sandy aquifer; sediment electrical resistivity; and annual precipitation. It might be expected that hydroperiod would be longer for clay pools than sandy pools because clay sediments can act as a seal to perch shallow ground-water and surface-water. Our data revealed the opposite to be true. Sandy residual sediments helped capture infiltration and direct this water along perched ground-water lenses or sheets to seasonal pools. This resulted in annual hydroperiods that were 115 days longer for sandy pools when compared to clay pools. The results suggest that the geologic substrate can be a major control on the duration of hydroperiod.

Evaluation of measurement errors by repeated pumping tests

Abstract A pumping test was replicated three separate times on a well near State College, Pennsylvania under controlled conditions to evaluate the magnitude of error and precision that can be achieved in measuring drawdown and specific capacity. Three step-drawdown tests were conducted and eleven different steps were used during each of the three tests. Pumping rates ranged from 60 to 109 gallons per minute. The grand mean and standard the deviation of the specific capacity values for the three tests are 11.825 gpm/ft and 0.722. Analysis of variance (ANOV) showed that the variation in specific capacity due to variations in pre-pumping water levels and pumping rates is significant.