# LINJ 1997 Groundwater

## Science Center Objects

The water quality in aquifer systems of the LINJ has been extensively investigated. The areal extent and number of these investigations, however, varies across the study unit. The fractured bedrock and valley-fill aquifers of the Piedmont and New England provinces in northern NJ have been the subject of fewer and more site-specific investigations, whereas, the unconsolidated sand and gravel aquifers of the Coastal Plain of LI and NJ have been the subject of numerous site-specific and regional water-quality investigations. Consequently, more is known about the geohydrology and water quality of the Coastal Plain aquifers.

### GW Background and Perspective

Numerous groundwater quality investigations undertaken as part of the Toxic Substances Hydrology Program in the Coastal Plain examined relations between land use and shallow groundwater quality in regionally extensive areas of LI and NJ (Barton and others, 1987; Eckhardt and others, 1989; Stackelberg, 1995; Eckhardt and Stackelberg, 1995; and Vowinkel and Battaglin, 1995). Eckhardt and others (1989) evaluated the relation between land use and groundwater quality in the surficial aquifer system of Nassau and Suffolk Counties on LI examining over 14,000 chemical analyses of samples from 903 wells collected between 1978-84. Results indicate that contamination from human activities has affected water quality in the surficial aquifer system. The occurrence of volatile organic compounds and pesticides was confirmed and statistically significant correlations between land use and these and other contaminants were established.

In a later investigation, Eckhardt and Stackelberg (1995), water-quality data from 90 monitoring wells screened near the water table beneath five areas of differing land use were compared. Results indicate that samples from undeveloped areas had the lowest and smallest range in concentrations of most human-derived constituents such as nitrate, boron, VOCs, and pesticides. Concentrations of these constituents in samples from three suburban areas and an agricultural area generally were intermediate to high and had the widest variation. Equations predicting the occurrence of contaminants near the water table were developed using logistic regression analyses of explanatory variables that characterize the type of land use and population density within a 1/4-mile radius of each of the 90 wells.

Vowinkel and Battaglin (in press) evaluated the effects of nonpoint sources of contamination on the quality of water in aquifers of the Coastal Plain in New Jersey. Water-quality data from over 1,000 wells sampled between 1980-89 for major ions, nutrients, trace elements, and organic constituents including dissolved organic carbon, phenols, volatile organic compounds, and pesticides were examined. Results indicate that nonpoint sources of contamination significantly affect the quality of shallow groundwater in aquifers of the Coastal Plain in New Jersey and that the distribution of contaminants are significantly related to patterns of land use.

The vulnerability of water withdrawn from public-supply wells to contamination by pesticides and volatile organic compounds has also been statistically and deterministically evaluated for aquifers in the Coastal Plain (Navoy, 1993; Vowinkel and others, 1994). Navoy (1993) utilized a finely discretized groundwater-flow model and flow-path simulation to demonstrate the vulnerability of Coastal Plain aquifers to non-point sources of contamination. Navoy (1993) applied a quantitative understanding of the groundwater flow system of the Potomac-Raritan-Magothy aquifer system in Gloucester County, NJ, to identify public-supply wells where contamination by VOCs was likely. Results from this analysis were consistent with available water-quality data and indicate that wells currently unaffected by contamination probably will be affected by contamination in the future and that the concentrations of VOCs in water from affected wells is likely to increase. The method of investigation utilized by Navoy (1993) demonstrates the (1) utility of a quantitative approach to understanding water-quality issues in water-supply aquifers, and (2) transferability of such methods to other locales where a finely discretized groundwater flow model and high-resolution land-use data are available.

Beginning in 1995, Safe Drinking Water Act regulations required the 626 large community water systems in New Jersey to monitor their 2,600 wells quarterly for 23 pesticides. As part of a 3-year study that began in October 1992, the USGS, in cooperation with the New Jersey Department of Environmental Protection (NJDEP), developed a geographic information system (GIS) data base to provide data on the vulnerability of water from public supply wells to contamination by pesticides (Vowinkel and others, 1994). Vulnerability was determined by using a numerical rating method based on information in the GIS data base. The information will be used by the State to determine the level of monitoring as a function of the vulnerability rating and to give waivers if a well is not vulnerable. The vulnerability of a well to contamination by pesticides is defined by (1) the sensitivity of the aquifer to contamination and (2) the intensity of pesticide use in areas where the aquifer is sensitive. Three variables were used to predict aquifer sensitivity: (1) location of a well relative to the outcrop area, (2) soil organic matter content, and (3) depth from the land surface to the top of the open interval of the well (top of screen for wells in unconsolidated sediments and top of the open hole for bedrock aquifers). Three variables were used to predict pesticide-use intensity near wells that are sensitive to contamination: (1) predominant land use near the well, (2) distance from the nearest agricultural area, and (3) distance from the nearest golf course.

Where information was available, well-construction characteristics and location were determined for 2,100 of the 2,600 public supply wells in New Jersey. These wells are located in three different aquifer types: (1) Coastal Plain unconsolidated sand and gravel deposits, (2) unconsolidated glacial-deposit sediments, and (3) fractured bedrock. Using the numerical rating method, each well was assigned to one of 9 vulnerability groups on the basis of its sensitivity and intensity ratings. It was determined that about 26 percent of all public supply wells are not vulnerable to contamination from human activities. All wells in this low vulnerability group are in the confined parts of Coastal Plain aquifers. About 4 percent of public supply wells are ranked in the high sensitivity and high intensity group. These wells belong to the high vulnerability group, are located within the unconfined parts of aquifers and are within or adjacent to agricultural land. The remaining 70 percent of wells were determined to be moderately vulnerable to contamination from human activities and are largely unconfined and within or adjacent to residential or agricultural land.

Predicted model results were validated by analyzing water samples for pesticides and nitrate from a subset of 90 public supply wells throughout New Jersey. Wells were chosen from each combination of vulnerability category and aquifer category. Pesticides were detected in 6 of the 90 wells sampled. Three of these wells were rated in the high sensitivity and high intensity group. The other three wells were rated as having either high sensitivity or high intensity. None of the six wells were rated as having low sensitivity to contamination by pesticides.

Groundwater studies like those for predicting nitrate and VOC contamination of groundwater on Long Island using regression analysis and for predicting pesticide contamination of groundwater in New Jersey using the vulnerability rating method benefit everyone. The NJDEP estimated that monitoring waivers for pesticides granted for wells will save taxpayers almost $5 million annually for a one-time cost of$0.7 million. In addition, consulting firms, other Federal, State, and county agencies, universities and the general public make requests daily for the water-use, water-quality, and hydrogeologic data resulting from these and similar studies.

### Groundwater Activities

All groundwater activities proposed in the FY96 workplan were accomplished as planned. This section describes these accomplishments and introduces proposed groundwater activities for FY97 and FY98.

HIP Plans as Modified by FY96 Conference Call
The present design for the LINJ GW effort encompasses:

(1) a new-urban Land-Use Survey in the NJ Coastal Plain (FY96-97) coupled with a regionally focused Flow-Path Survey in the same area (FY96-97),

(2) a SUS of the valley-fill aquifers in northern NJ (FY 97-98),

(3) a localized surface-/groundwater interactions Flow-Path Survey (FY98) and/or a water-supply relevant investigation of aquifer vulnerability on LI (FY98), and

(4) retrospective SUS and/or LUS (FY 98) of areas not covered above in as much as possible.

It is not anticipated that all items identified above will be accomplished by the end of FY98. The likelihood and scope for those items proposed for FY98 will be determined pending the completion of FY97 activities and the evaluation of data collected by GW and SW efforts during FY97.

### Study-Unit Surveys

Study-unit surveys are to be conducted in areas that have limited groundwater quality data relative to other parts of the study unit, where drilling new wells is not practical, and where groundwater provides an important source of potable water. Primary objectives for study-unit surveys are to (1) provide a broad assessment of groundwater quality conditions in regionally extensive aquifers utilized for public supply, (2) identify those constituents causing the most prevalent water-quality concerns, and (3) to the extent possible, describe the occurrence and distribution of select constituents in relation to land-use patterns, geohydrology, soil types, and other natural and anthropogenic factors. The need to broadly assess water-quality conditions in major hydrogeologic units will vary among different study units depending on the availability and quality of existing data and the size and complexity of the groundwater systems. Study units with substantial water-quality data for the highest priority subunits may wish to allocate a larger portion of study resources towards spatially focused studies designed to increase the understanding of the sources and processes controlling the movement and fate of contaminants in groundwater systems.

Available water-quality data for the primary LINJ groundwater subunits have been compiled and screened to prioritize subunits in most need of study-unit surveys and to identify subunits in which analysis of existing data or sampling of existing wells may accomplish study-unit and/or land-use survey objectives. This section briefly describes the distribution of available water-quality data and proposed plans to conduct a study-unit survey in our highest priority subunit.

Accomplishments in FY96

Substantial amounts of water-quality data exist for groundwater subunits in the Coastal Plain province of the LINJ study unit. The relation between groundwater quality and land use was the focus of investigations undertaken in this province as part of the Toxic Substances Hydrology Program during the 1980's and early 1990's. Results from these and other investigations have been summarized in a retrospective lay-reader type publication tentatively entitled How are people affecting groundwater quality in New Jersey and Long Island' by Clawges and others (in review). The scope and major findings from investigations of regional water quality in the Coastal Plain Province are summarized in the GW background section at the end of this workplan. Study-unit surveys are not currently planned for subunits in the Coastal Plain due to the existence of extensive water-quality data sets and previous investigations which evaluated the relation between groundwater quality and land use in this province.

Considerably less regional-scale water-quality data exists for other groundwater subunits in the LINJ study area (Table GW-1). This is especially true for the valley-fill and fractured-bedrock aquifer systems of northern NJ (NE NJ) which are utilized extensively for public supply. Based on liaison input, review of existing literature, and available water-quality data, the LINJ team has given the valley-fill aquifer systems the highest priority in terms of need for a study-unit survey.

Proposed Work in FY97

A study-unit survey is planned for the valley-fill aquifers of the Piedmont/New England provinces (NE NJ) due to a relative scarcity of regional-scale water-quality data (Table GW-1), an increasing demand for public supply, and a highly vulnerable hydrogeologic setting. About two thirds (nearly 4 million people) of the population of NJ resides within this area and about 59% of purveyor-supplied water is groundwater. In addition, more than 71,000 domestic wells supply about 9% of the residents with water (NJDEP, 1993). North of the line of glaciation, valleys are generally filled with stratified and unstratified glacial and lake deposits. Where the valley-fill deposits are thick and permeable they provide significant quantities of water. For example, in Essex County about 81% of the pumped groundwater is from valley-fill aquifers (NJDEP, 1993). Many of the valley-fill aquifer systems are in highly populated/developed areas and many others are in areas currently experiencing rapid growth. Due to their highly permeable nature and hydraulic connection to land surface, these aquifers are highly susceptible to contamination introduced at or near land surface. Retrieval of available water-quality data for groundwater subunits of the LINJ study unit indicates that relatively little data exists for the valley-fill aquifer systems (Table GW-1). For these reasons, the valley-fill aquifers subunit will be targeted for a study-unit survey.

The study-unit survey will be designed to sample groundwater under unconfined conditions across a range of land-use environments. Between 30 and 45 wells screened within valley-fill sediments will be randomly selected using a grid-based random site selection program (Scott, 1990). To the extent possible, only observation and monitoring wells screened within 100 feet of land surface will be utilized. Sampling is scheduled for FY98.

The standard suite of chemical constituents and analytical methods identified by the NAWQA program will be utilized in this study-unit survey. Specifically, sampling will include field parameters, major ions (SH2750), nutrients (SH2752), pesticides (SH2001 and SH2050), volatile organic compounds (SH9090), and dissolved arsenic. Pesticide schedules 0079 and 1321 are potential add-ons, the need for which will be evaluated before sampling takes place. Quality-assurance samples will be collected at fifteen percent of the sites.

Land-Use Surveys

Due to the highly urban nature of the LINJ study unit a land-use survey focusing on urban land was proposed for FY96. Primary objectives for urban land-use surveys are to (1) assess the concentrations and distribution of water-quality constituents in recently recharged groundwater associated with areas of mixed residential and commercial land-use less than 25 years old, and (2) define the natural and anthropogenic factors most clearly associated with observed water-quality conditions in these land-use settings. This section describes the urban land-use survey accomplished in FY96 and our extension of the urban land-use survey to address liaison committee concerns and to coordinate with the comprehensive urban investigation. Because we are proposing to initiate a study-unit survey and complete a flow-path survey during FY97, a new land-use survey is not proposed for FY97.

Accomplishments in FY96

Seventy-two shallow monitoring wells were located, installed, and sampled during summer and fall 1996 (FY96 and FY 97) in the Glassboro region (Fig. GW-1) of the NJ CP. Well sites were distributed throughout all major land-use settings within the study area (Fig. GW-2). Thirty wells were located in newly developed (<25 years) urban areas (lus new urban NJ1). The additional 42 wells (14 in older (>25 years) urban areas, 15 in agricultural areas, and 13 in undeveloped areas) were installed as part of the flow path study (fps ag/urban NJ1) to address concerns raised by our liaison committee and to coordinate data-collection activities with the comprehensive urban investigation (described in separate workplan). The LINJ liaison committee strongly recommended the LINJ go beyond identifying issues in these areas and design a program that will address vulnerability of public and domestic water supplies to contamination from urban, residential, and agricultural land-use activities; thus, our new-urban land-use survey was extended to include a representative number of sites from all major land-use settings. This extension of the land-use survey provides water-quality data from other majorland-use settings and allows for a more rigorous comparison of the effects of new urban land use. The LINJ liaison committee also recommended we identify sources of contaminants and processes which most significantly affect the transport and fate of select contaminants as they move through major compartments of the hydrologic cycle. Thus, a comprehensive urban investigation workplan was developed which builds upon the core LINJ data-collection activities and addresses the major source, transport, and fate issues. The 72 well extended land-use survey represents an important component of the flow path study and comprehensive urban investigation by providing occurrence and distribution data at the water-table from all major land-use settings in our study area. The 72 well network constitutes our initial efforts at a regional-scale flow-path survey as described in the next section entitled Flow-Path Surveys'. The background and initial water-quality results for the 72 well survey is as follows.

The study area depicted was targeted for an extended land-use survey as part of the core LINJ workplan and comprehensive urban investigation for multiple reasons. The study area falls within the NJ State's Critical Water Supply Area #2 where serious water-supply problems have been identified by the NJ Department of Environmental Protection (NJDEP). Due to severe water-level declines in the confined Potomac-Raritan-Magothy aquifer system in this area the NJDEP has restricted further withdrawals from this aquifer system and recommended increased withdrawals from the surficial Kirkwood-Cohansey aquifer system, thereby placing additional stress on the surficial system to meet current and future water-supply demands. Currently, there are over 400,000 residents within the urban land-use study area and population values are expected to increase 20% by the year 2010 due to increasing suburban development (NJDEP, 1993). The surficial aquifer system in this area is comprised of highly permeable unconsolidated sands and gravels; thus, the system is highly vulnerable to contaminants introduced at or near land surface. The study area is also within the Philadelphia metropolitan area which is an EPA non-attainment region with respect to air quality, therefore, gasoline oxygenated with MTBE has been used year-round as mandated by the Clean Air Act Amendments of 1990.

Our new urban land-use survey consists of 30 wells and was designed following guidelines established by Squillace and Price (1996), as were the other 42 wells. Installation and sampling of all 72 wells in the extended land-use survey was accomplished using protocols and procedures described in Lapham and others (1995) and Koterba and others (1995). Table GW-2 provides a summary of the major land-use settings targeted as part of this extended survey, the number of wells installed per land-use setting, and the number of wells sampled for various constituents.

Potential well sites within each land-use setting were randomly selected using a grid-based random site selection program (Scott, 1990). Each well is constructed of 2-in. diameter PVC and was installed by hollow-stem auger. Wells were generally screened over a 2-foot interval about 10 feet below the water table. Continuous split-spoon samples of the unsaturated zone were collected during well installation at 20 new-urban sites and all of the old urban, agricultural, and undeveloped sites. These samples will be analyzed for sediment size, organic carbon content, soil moisture and pH, and the presence or absence of local finer-grained units. These characteristics are important because they may effect the movement of certain organic compounds through the unsaturated zone. All wells were sampled for the standard suite of chemical constituents identified by the NAWQA program including field parameters, major ions (SH2750), nutrients (SH2752), pesticides (SH2001/2050, SH1321, and SH0079), and volatile organic compounds (CM9090). Quality assurance samples were collected at 24% of the sites.

Potential anthropogenic sources of contamination are being documented within a 500m radius of each well and stored in a GIS format to facilitate spatial analysis (LU/LC pilot). Data to be stored in GIS format include (1) the location of gas stations, dry cleaners, TRI sites, land fills, and other point sources, (2) local pipeline, sewer, and road networks, and (3) detailed representations of current land-use patterns and population density.

### Flow-Path Surveys

Primary objectives for flow-path surveys are to (1) characterize the spatial and temporal distributions of water quality in relation to groundwater flow in particular land-use settings, and (2) further the understanding of natural processes and human influences in these settings which affect the evolution of groundwater quality along flow paths (Gilliom and others, 1995). The primary concern expressed by our liaison committee was to address the vulnerability of public and domestic water supplies to contamination from all relevant land-use settings. This section describes the initiation of a regionally-focused flow-path survey designed to address NAWQA objectives and liaison committee concerns in an area with critical water-supply issues.

Evaluating the source, transport, and fate of contaminants from a water-supply perspective requires occurrence and distribution data for contaminants of interest and the development of three-dimensional mathematical models of groundwater flow and reactive chemical transport over temporal and spatial scales relevant to water-supply planning and aquifer development. The LINJ study team has expanded the scope of NAWQA flow-path surveys to incorporate modeling and other methods to further the understanding of current and future distributions of contaminants in a water-supply relevant aquifer system.

The 72 well monitoring network installed in the Glassboro region during FY96 was designed to randomly sample recently recharged groundwater. The network constitutes both an urban land-use survey and our initial efforts at a regionally focused flow-path survey by providing detailed information on the occurrence and distribution of specific contaminants at the water table from all major land-use settings in the study area. This survey, when linked with land use and other data, provides the basis for estimating loadings of specific compounds to the water table as the starting point (input) to the groundwater flow and transport model.

A three-dimensional groundwater-flow model of the surficial Kirkwood-Cohansey aquifer in the study area was developed in FY96. The model (1) quantifies the rate and path of groundwater movement through the aquifer system under current and projected water-use scenarios, (2) quantifies the time required for water to move along given flow paths from the water table to eventual discharge at a stream or water-supply well, and (3) establishes a basis for predicting the rate of contaminant movement along various flow paths. This flowpath analysis will be used to evaluate existing wells and future well locations for sampling of groundwater of at greater depth (different ages and different land use sources) as part of the flow path study. Site selection, installation, and sampling are proposed for FY97. The number of sites and analytical schedules in this deeper network will be very focused as determined from an analysis of water-quality data from the 72 well network and modeling. Chemical transport modeling to determine the relevance of compound occurrence and importance to water supply in this surficial aquifer system is proposed for FY98.

Pending the timely completion of our efforts in the Glassboro region, the LINJ groundwater team is contemplating a similar modeling flow path investigation on LI for FY98. The likelihood and scope of this work will not be determined until current work in the Glassboro region is nearer completion. In addition, pending the completion of a proposed VOC reach study of several LI streams, the groundwater team is contemplating an investigation into the role of shallow groundwater as a source of VOCs in LI streams. The likelihood and scope of this work will not be determined until results from the reach study are available. The following discussion provides background information and preliminary concepts for these potential activities.

The nearly 2.6 million residents of Nassau and Suffolk Counties on LI are entirely reliant on groundwater for public supply. The surficial aquifer system has been abandoned as a source of public supply in much of the two-county area due to documented contamination. The underlying aquifer system is hydraulically connected to the surficial system and, thus, is vulnerable to contamination introduced into the surficial aquifer. Both aquifers are composed of unconsolidated sands and gravels which are generally highly permeable.

A three-dimensional finite-difference model with particle-tracking capabilities has been developed for the groundwater system of LI. This model can be utilized to quantify the pattern and rate of groundwater flow under both natural and stressed conditions. In addition, extensive water-quality data sets exist for Long Islands groundwater resources. For instance, Nassau and Suffolk Counties maintain monitoring well networks of over 400 and 300 wells, respectively. These networks are sampled annually for nutrients, major ions, trace elements, and VOCs. In addition, Suffolk County samples for pesticides. The possibility exists to couple the groundwater flow model and water-quality data sets to address the vulnerability of public-supply wells to contamination introduced at the water table. By estimating contaminant loading rates for specific land-use settings, utilizing available data to document the current distribution of contaminants, and utilizing the groundwater flow model to establish patterns and rates of flow through LI's aquifer systems, the potential for future contamination of existing public supply wells can be estimated.

Streamflow on Long Island is derived primarily from shallow groundwater, thus, groundwater contamination may be a source for contaminants found in surface-water bodies. Data collected by the Suffolk County Department of Health confirms the occurrence of VOCs in both surface- and groundwater resources of LI. A factsheet describing the presence and distribution of VOCs in Suffolk Counties streams is in progress and a publication describing the occurrence of MTBE in Suffolk Counties surface- and groundwater bodies has been accepted for presentation at the American Chemical Society's National Meeting.

To further evaluate the source(s) of VOCs in LI's streams, the LINJ SW team plans a VOC reach study of streams in two or three watersheds of differing land use. To compliment the VOC reach study and evaluate shallow groundwater as a potential source of VOCs, the LINJ GW team may conduct a flow-path survey. The location and scope of this flow-path survey would be determined pending results of the reach study. The possibility exists to utilize the groundwater flow model to define the extent and rate of movement along flow paths that discharge to streams in the reach study. Once defined, a series of groundwater transects could be established to characterize the spatial distribution of contaminants and identify the factors most significantly affecting the evolution of water-quality along these flow paths. Detailed sampling at the surface- groundwater interface would allow for evaluation of groundwater as a source of VOC contamination to LI's streams.