DESIGN AND ANALYSIS OF TRACER TESTS TO DETERMINE EFFECTIVE POROSITY
AND DISPERSIVITY IN FRACTURED SEDIMENTARY ROCKS, NEWARK BASIN, NJ

PI: Herb Buxton, Co-PI: C. Welty
Project Chief: Glen Carleton

Field-scale pumping and tracer tests were conducted in 1994-95, at a site in Hopewell Township, Mercer County, NJ for the purpose of characterizing hydraulic and transport properties of the fractured sedimentary rock. Prior to this work, little quantitative information on solute-transport properties of this aquifer had been collected. The study was conducted by the U.S. Geological Survey in cooperation with the New Jersey Department of Environmental Protection.

The site is located in the Newark Basin, an elongate (210 by 55 km), northeast-southwest trending fault trough filled with late Triassic and early Jurassic fluvial and lacustrine sediments and igneous intrusions, that is part of the Piedmont Physiographic Province. The site is underlain by the Passaic Formation of Late Triassic age, an important aquifer in New Jersey and Pennsylvania, consisting of red arkosic mudstones, siltstones, and fine-grained sandstones. Bedding planes strike approximately east-west and dip moderately to the north in the
vicinity of the wells tested. The two dominant fracture sets are bedding plane partings and east-west striking structural fractures that dip steeply to the south.

A pre-existing network of 13 wells at the site includes a central well (well 1), well 6 located 30.5 m downdip (north), wells 2 and 10, located 91.4 and 183 m, respectively, approximately along strike (west), and well 5, located 92.7 m downdip. The wells are constructed with 6 m of steel surface casing and are about 46 m deep, except for well 6, which was drilled to 61 m in order to penetrate the same bedding planes intersected by well 1. A full suite of geophysical logs -- including gamma, electric, electromagnetic conductance, caliper, fluid temperature and resistivity, video, acoustic borehole televiewer, and heat-pulse flow meter -- was collected and interpreted as part of this study. The geophysical logs were used to determine location and orientation of fractures and to construct lithologic sections that correlated producing zones on the basis of geologic characteristics.

A nine-day aquifer test was conducted in October 1994 by pumping water from a withdrawal well (well 1) at 108 L/min (+/- 2 %), and recording the hydraulic heads as a function of time in the withdrawal well and 14 observation wells. The drawdown curves differ significantly in the strike and dip directions, indicating that the site is characterized by anisotropic hydraulic conductivity. The analytical method of Hsieh and Neuman (WRR, 1985, case 4) and Hsieh and others (WRR, 1985) was used to make preliminary calculations of the aquifer hydraulic properties. The analysis yielded principal values of hydraulic conductivity of 6.4, 0.30, and 0.0043 m/d, and a value of specific storage of 9.2 x 10-5 m-1. The maximum principal direction of hydraulic conductivity was nearly aligned with strike and approximately parallel to land surface; the minimum value was perpendicular to land surface (across the bedding planes). The analytical results were used as a guide for calibrating a three-dimensional numerical model consisting of 9 layers aligned with the bedding planes, to the head data. The best-fit values of hydraulic conductivity from the numerical modeling were 7, 3, and 4 x 10-5 m/d for the strike, dip, and normal-to-bedding-plane directions, respectively.

Three non-recirculating doublet tracer tests were conducted at well spacings of 30.5 m, 91.4 m, and 183 m in the 40-m-long open boreholes, using pulse injections of bromide. Water was injected into well 6, 2, or 10 and withdrawn from well 1 at flow rates of approximately 120 L/min for each test. Well 5 was used as a source of water for injection because it was believed to be isolated enough hydraulically from the pumped well so as to not interfere with the hydraulic regime of the tracer tests. Water withdrawn from well 1 was discharged to the pond. Bromide concentrations were measured in water withdrawn from well 1.

An analytical solution of for a doublet tracer test was applied to the rising limbs and peaks of collected bromide concentration versus time (breakthrough) data to obtain estimates of dispersivity and effective porosity. Calculated values of dispersivity for each test were: 4.6 m for the well 6 - well 1 test (a scale of 30.5 m); 10.1 m for the well 2 - well 1 test (a scale of 91.4 m), and 12.8 m for the well 10 - well 1 test (a scale of 183 m). These dispersivity estimates are considered to be approximate because the analytical solution assumes constant dispersivity, homogeneous and isotropic media, and an infinite domain. Effective porosity values obtained from the analyses ranged from 3.7 x 10-4 to 3.0 x 10-3. The tracer mass recoveries were lower than expected, which may be explained in part by flow being diverted toward well 5 and along pathways not simulated. A two-dimensional, finite-element transport model was calibrated to the hydraulic data and to the well 10 - well 1 (183 m scale) tracer test data, for the purpose of incorporating effects of anisotropy and boundaries. The value of dispersivity obtained from the numerical modeling was identical to that obtained from the analytical modeling (12.8 m); the best-fit value of effective porosity was 1.2 x 10-3.

Test results indicate that dispersivity increases with scale at the site. Whether the value of dispersivity of 12.8 m could be considered a constant, asymptotic value, or whether the concept of Fickian transport is valid at this site could only be ascertained if further tracer tests were conducted at larger scales. The values of dispersivity found are greater than most reported for porous media sites at similar scales, but typical of more heterogeneous fractured rock environments.

Based on the shape of the breakthrough curves, matrix diffusion did not appear to be an important process, but this could be because of the rapid transport typical of forced-gradient doublet tests. The influence of matrix diffusion could be ascertained using natural-gradient tests that more realistically mimic natural flow conditions or tests employing multiple tracers having different molecular diffusion coefficients.

The study site offered a unique opportunity to conduct tracer tests at multiple scales for a relatively low cost. Existing information on dispersive properties of fractured sedimentary rock aquifers is scant (National Research Council, 1996); this study served as a first step in characterizing this type of hydrogeologic environment in New Jersey. The results from this study indicate that continuum-based porous medium models can be used reasonably well in application to hydraulic and tracer tests in this type of fractured sedimentary rock.

Potentially fruitful future work at this site could include: (1) conducting tracer tests at larger scales to address the question of Fickian transport; (2) designing and running appropriate tests to assess the importance of matrix diffusion, and (3) collecting and analyzing further heat-pulse flowmeter data and relating a quantitative geostatistical analysis of small-scale hydraulic conductivity data from the flowmeter test to the previously collected large-scale dispersivity and hydraulic conductivity information.

Project Duration: 10/1/93 - 7/30/96
Funding Source (to USGS): NJ Dept. of Environmental Protection

Publication:

Carleton, G.B., C. Welty, and H. Buxton. Design and analysis of tracer tests to determine effective porosity and dispersivity in fractured sedimentary rocks, Newark Basin, NJ, USGS Water Resources Investigations Report 98-4126A, 80 p., 1999.