Evolving Conceptual Models and Monitoring Well Reconstruction in the Passaic Formation in New Jersey Abstract

Transcription

1 Evolving Conceptual Models and Monitoring Well Reconstruction in the Passaic Formation in New Jersey John N. Dougherty and Andrea Soo, CDM. Robert M. Alvey, U.S. Environmental Protection Agency Abstract This paper will describe a monitoring well reconstruction program designed to test and support a revised conceptual groundwater flow model for a Federal Superfund site in New Jersey underlain by the Passaic Formation. Well reconstruction was undertaken to produce a monitoring well network capable of providing three-dimensional groundwater elevation, and water quality data from the fractured bedrock aquifer The monitoring wells were initially constructed in the 1980s during the remedial investigation of the site. The original site conceptual model envisioned the site as being underlain by overburden and weathered bedrock unit then an anisotropic, bedrock water table aquifer; and then a semi-confined or confined anisotropic bedrock aquifer. Fractures and joints were postulated as important pathways for groundwater flow. The major axis of anisotropy was parallel to bedrock strike of north 65 degrees east (N65E). The bedrock wells were cased into competent bedrock and open, 6-inch boreholes were extended to depths ranging from 88 to 250 feet. Depths of casing ranged from 19 to 200 feet. The site conceptual model was later updated, based on published sources, and envisioned as an anisotropic, heterogeneous, multi-layer aquifer system where zones of relatively high permeability were separated by zones of relatively low permeability. The zones of high permeability correspond to bedding planes in the Passaic Formation which strikes N65E and dips to the northwest at 17 degrees. In this model the long open intervals in the monitoring wells may intercept multiple conductive zones in the aquifer. One consequence of this is that the water level and water quality data from the wells are averaged over the open interval. To allow measurement of water level and water quality at specific, relatively short intervals in the aquifer and thereby provide greater resolution in assessing groundwater flow and contaminant migration, a monitoring well reconstruction program was planned and executed. Available core logs, driller s notes, borehole television and packer testing data were utilized to determine where to place screened intervals in each borehole. A borehole television and packer testing program was planned for wells that had not been previously tested. The borehole television program was conducted first and the results were used to select locations for packer testing in each borehole. Packer testing was then conducted to collect groundwater samples and hydraulic data from specific 20-foot intervals in the aquifer. The hydraulic data were analyzed to estimate the hydraulic conductivity of each interval. The results of the packer testing conducted at all the wells were reviewed and 20-foot screened intervals were selected to complement one another. For example, at a location where three wells were located in proximity, screened intervals were chosen in shallow, intermediate, and deep conductive zones to provide vertical coverage of the aquifer. After the wells were reconstructed and surveyed, two rounds of groundwater samples and three rounds of water level measurements were collected. These data were used to develop isoconcentration contour maps and potentiometric surface maps. Since data were now available in three dimensions, a series of cross sections were prepared. The cross-sections were drawn perpendicular to bedrock strike (parallel to dip). The water level elevation and contaminant concentrations were posted adjacent to the mid-point of each screened interval. These data were then contoured, taking into account the dip of the beds and variation in permeability. Next, an elevation was chosen at which to slice through the aquifer to produce a plan view. Data were then taken from each cross section at the selected elevation and posted in plan view along the cross section line. These data were then contoured to produce a plan view groundwater contour and isoconcentration contour maps. The well reconstruction program produced a monitoring well network capable of providing three-dimensional groundwater elevation, and water quality data from the fractured bedrock aquifer. These data will be used to analyze and assess the performance of a groundwater remediation system planned for the site. 553

2 Introduction The Rocky Hill Municipal Well Field (RHMW) and Montgomery Township Housing Development (MTHD) Superfund Sites are located, respectively, in the Borough of Rocky Hill and in Montgomery Township, Somerset County, New Jersey, along the west bank of the Millstone River (Figure 1). The 2-acre RHMW Site is located east of New Jersey State Route 206 and just south of New Jersey State Route 518. It is centered on the Rocky Hill Municipal Well #2 (RHMW #2 on Figure 1). The MTHD Site consists of 71 approximately one-acre home sites located in Montgomery Township, New Jersey. The homes are on Montgomery Road, Sycamore Lane, Robin Drive, Oxford Circle, and Cleveland Circle, and are east of New Jersey Route 206 and north of Route 518 (Figure 1). Groundwater at each site is contaminated with volatile organic compounds (VOCs) and in particular trichloroethene (TCE). The sites are being addressed jointly. In 1979 TCE contamination was first detected in groundwater samples from the municipal well and in about 50% of the 71 domestic wells in the development. The two sites, which are adjacent to one another, were included on the National Priorities List (NPL) in In 1983, the Borough of Rocky Hill installed an air stripper on the municipal well and resumed using the well to provide potable water to about 1,000 customers in the Borough. The municipal well, constructed in 1936, continues to supply the Borough up to the present. In 1985 the New Jersey Department of Environmental Protection (NJDEP) initiated a Remedial Investigation/Feasibility Study (RI/FS) which was completed in A Record-of-Decision (ROD) was signed for the site in The ROD required residents to hook up to the public water supply and called for the installation of a pump and treat system to remediate groundwater at the site. In 1993 NJDEP completed a 65% design of a groundwater treatment system but did not proceed with construction. At that point, the U. S. Environmental Protection Agency (EPA) became the lead agency for the sites. In 2001, EPA began the process of finalizing the remedial design. Construction of a groundwater pump and treat system began in The sites are located in the Piedmont Physiographic province. Topography are the site ranges from a high of about 150 feet above mean sea level (msl) near the municipal well to about 20 feet near the Millstone River to the east and at Beden Brook to the north. Conceptual Model Developed During the Remedial Investigation The original site conceptual model envisioned the site as being underlain by overburden, up to 30 feet thick composed of clay, silt, and rock fragments, a weathered bedrock unit, then an anisotropic, bedrock water table aquifer; and then a semi-confined or confined anisotropic bedrock aquifer (Woodward Clyde Consultants 1988). Vertical fractures parallel and perpendicular to strike and occasional steeply dipping joints that cross cut the vertical fractures were postulated as important pathways for groundwater flow. In this model, bedding planes were not considered important groundwater flow pathways. This conceptual model was based on the work of Herpers and Barksdale (1951), Vecchioli et al. (1967), and Vecchioli et al. (1969). The major axis of anisotropy was parallel to bedrock strike of north 65 degrees east (N65E). Bedrock dip was estimated at 12 to 13 degrees northwest. During the RI a total of 30 monitoring wells were installed in shallow and deep pairs. In general, the wells were installed along a north-south trending path perpendicular to strike. The shallow wells were installed to monitor the weathered bedrock zone and ranged in depth from 20 to 82 feet. All shallow wells were screened. The deep wells were installed to monitor groundwater in the competent rock and ranged in depth from 100 to 250 feet. All the deep wells were completed as open holes. After the wells were completed, water level data was collected. Figure 2 shows a potentiometric surface for the deep wells prepared using this data. Using the conceptual model described above. The data showed that water level elevations in the shallow wells were higher than in the adjacent deep well. This was interpreted as evidence of a water table aquifer at depths ranging from 3 to 40 feet below ground surface (bgs). 554

3 Conceptual Model Used During the GW Treatment System Design in 1993 During the initial design of a groundwater treatment system at the site in 1993, the conceptual model of groundwater flow in the Passaic Formation beneath the sites was revised based on the work of Michalski (1990), the mapping of Parker and Houghton (1990), and the hydraulic conductivity and TCE data derived from packer testing conducted in The new model envisioned the aquifer as an anisotropic, heterogeneous, multi-layer aquifer system where zones of relatively high permeability are separated by zones of relatively low permeability. Cross sections A-A and B-B (Figures 3 and 4) are drawn through the conceptual model parallel to dip. The shaded areas on the figures represent zones of increased permeability caused by dipping bedding plane partings in the Passaic Formation.) The locations of these cross sections are shown in Figure 5. Michalski s work showed that groundwater flow in the Passaic Formation is strongly influenced by bedding plane orientation with the highest permeabilities being observed parallel to bedding strike. Mapping by Parker and Houghton (1990) indicates that the Passaic Formation in the vicinity of the site has undergone little deformation and that bedding strikes north 65 degrees east (N65E) and dips to the northwest at 17 degrees towards the Hopewell Fault which is located about 5 miles northwest of the site. A diabase sill outcrops to the south of the site and forms the Rocky Hill. The packer testing data collected in 1992 showed a background of relatively low hydraulic conductivity with isolated zones that exhibited hydraulic conductivity values one to two orders of magnitude greater than the background values. These zones of relatively high conductivity correlated reasonably well along strike. Hydraulic conductivity was observed to decrease with depth which is consistent with increasing lithostatic pressure reducing the size of openings along bedding plane partings. While the conceptual model was revised, the monitoring wells were not reconstructed. In 1998, the EPA conducted geophysical logging and borehole televiewer logging in monitoring wells MW-7D, MW-9D, MW-10D, MW-11D, MW-15D, MW-16D, MW-20D, and FD-1D (Weston 1998). The strike and dip data obtained from the borehole televiewer is consistent with the strike of N65E and dip of 17 degrees northwest. Conceptual Model Used During the Revised GW Treatment System Design, The current conceptual mode of the site envisions the aquifer as a leaky multi-unit aquifer system (LMAS) (Michalski and Britton, 1996). In this conceptual model groundwater movement through the Passaic Formation is controlled by a combination of movement along bedding planes and leakage between bedding planes through vertical joints. In 2001 and 2002, the EPA conducted a pre-design investigation in preparation for completing the groundwater treatment system. A prime goal of the pre-design investigation phase was to reconstruct the existing monitoring wells at the site based on the new conceptual model and to allow them to be used to test the model. The reconstructed wells allow measurement of water level and water quality at specific, relatively short, 20-foot, intervals in the aquifer and thereby provide greater resolution in assessing groundwater flow and contaminant migration. Available core logs, driller s notes, borehole television and packer testing data were utilized to determine where to place screened intervals in each borehole. Additional borehole television and packer testing was planned for wells that had not been previously tested. The borehole television program was conducted first and the results were used to select locations for packer testing in each borehole. Packer testing was then conducted to collect groundwater samples and hydraulic data from specific 20-foot intervals in the aquifer. The hydraulic data were analyzed to estimate the hydraulic conductivity of each interval. The results of the packer testing conducted at all the wells were reviewed and 20-foot screened intervals were selected to complement one another. For example, at a location where three wells were located in close proximity, screened intervals were chosen in 555

4 shallow, intermediate, and deep conductive zones to provide vertical coverage of the aquifer. Figures 3 and 4 show the locations of the screened intervals selected during the pre-design investigation phase. The new data generated during the pre-design investigation, such as permeability and TCE data from packer testing and water level data were also incorporated into the numerical flow model. The groundwater flow model was used to evaluate pump and treat system without reinjection of groundwater (reinjection was part of the original design). After the wells were reconstructed and surveyed, two rounds of groundwater samples (February and August 2002) and three rounds of water level measurements (February, March, and August 2002) were collected. These data were used to develop isoconcentration contour maps and potentiometric surface maps. Since data were now available in three dimensions, a series of cross sections were prepared. The cross-sections were drawn perpendicular to bedrock strike (parallel to dip). The water level elevation and contaminant concentrations were posted adjacent to the mid-point of each screened interval. These data were then contoured, taking into account the dip of the beds and variation in permeability. Next, an elevation was chosen at which to slice through the aquifer to produce a plan view. Data were then taken from each cross section at the selected elevation and posted in plan view along the cross section line. These data were then contoured to produce a plan view groundwater contour and isoconcentration contour maps. Figure 6 shows the February 2002 potentiometric surface and Figure 7 shows the February 2002 TCE data at the 50 foot elevation in the aquifer. The 50 foot elevation was chosen because contamination is most extensive at this elevation and concentrations are relatively high. Prior to reconstruction of the monitoring wells it was not possible to evaluate groundwater flow and water quality at different elevations in the aquifer. The revised, updated conceptual model was used in designing the monitoring well network for the groundwater treatment system. The original monitoring well network was oriented along dip in a north-south path. Based on the revised conceptual model, additional monitoring wells were proposed for installation along strike, both upgradient and down gradient, with respect to the two groundwater pump and treat systems (Figure 8). References Herpes and Barksdale. 1951, Preliminary Report on the Geology and Ground-Water Supply of the Newark, New Jersey Area, Special Report 10, New Jersey Department of Conservation and Economic Development, Division of Water Policy and Supply. Michalski, A Hydrogeology of the Brunswick (Passaic) Formation and implications for ground water monitoring practice. Ground Water Monitoring Review, vol. X, No. 4, p Michalski, A., and Britton, R The role of bedding fractures in the hydrogeology of sedimentary bedrock - Evidence from the Newark Basin, New Jersey. Ground Water, vol. 35, No. 2, p Parker, Roland A. and H.F. Houghton Bedrock Geologic Map of the Rocky Hill Quadrangle, New Jersey; U.S. Geological Survey, Open-File Report Vecchioli, J. 1967, Directional Hydraulic Behavior of a Fractured Shale Aquifer in New Jersey, Proceedings from Symposium on the Hydrology of Fractured Rock, International Association of Scientific Hydrology. Vecchioli, J., L.D. Carswell, and H. F. Kasabach 1969, Occurrence and Movement of Groundwater in the Brunswick Shale at a Site Near Trenton, New Jersey, United States Geological Survey Professional Paper 650-B, pp B154-B157. WESTON Final Report, Hydrogeologic Investigation Rocky Hill Municipal Well Superfund Site, Former Fifth Dimension Facility, Montgomery Township, New Jersey, prepared by Roy F. Weston, Inc. under the REAC Contract for the USEPA, Environmental Response Team Center. September Woodward Clyde Consultants 1988, Remedial Investigation/Feasibility Study for Montgomery Township Housing Development/Rocky Hill Municipal Well Field Site, Volume I, Remedial Investigation, Woodward-Clyde Consultants, April

5 Biographical Sketches John N. Dougherty (lead/presenting author) CDM Raritan Plaza I Raritan Center Edison, New Jersey Office: Facsimile: doughertyjn@cdm.com Mr. Dougherty is a hydrogeologist with 20 years experience in the environmental consulting. He has extensive experience in many aspects of site investigation and has designed and supervised the installation of numerous monitoring wells at Superfund sites around the United States. Mr. Dougherty holds a B.S. Degree in Geosciences from The Pennsylvania State University. Andrea Soo CDM 993 Old Eagle School Road Suite 408 Wayne, PA Office: (610) Facsimile: (610) sooa@cdm.com Mrs. Soo is a geologist with 5 years experience conducting remedial investigations, feasibility studies, remedial design and operation and maintenance of treatment systems through project planning, field work, data interpretation and reporting stages. She has worked on over 20 sites on the National Priority List. Robert M. Alvey, P.G. U.S. Environmental Protection Agency Emergency Remedial Response Division, 18 th Floor 290 Broadway New York, New York Office: (212) Facsimile: (212) alvey.robert@epamail.epa.gov Robert M. Alvey is a geologist with the USEPA Region 2 Emergency Remedial Response Division's Program Support Branch. He is responsible for the hydrogeologic investigations for a variety of Superfund sites in New York, New Jersey, and Puerto Rico. Rob is a licensed geologist and received BS and MS degrees in geology from Rensselaer Polytechnic Institute in

6 558

7 559

8 560

9 561

10 562

11 563

12 564

13 565