# LINJ Meeting Summary April 1995

## Science Center Objects

04/20/95-Meeting Summary

Dear liaison committee member,

Subject: Summary of the April 20, 1995 NAWQA liaison meeting in Edison, NJ.

Plans for the surface-water (SW) and groundwater (GW) activities for the next 6-8 months were presented and discussed at this third NAWQA liaison meeting. Plans for those activities that will support the study design included tasks to:

• Carry out the study unit (SU) stratification and other GIS activities o Prepare a tabular listing of the more useful available data bases Assess the cooperative streamflow and water quality networks for potential study sites
• Plan and carry out the field reconnaissance and site selection for SW monitoring sites
• Plan and carry out the field reconnaissance and site selection for the GW study-unit survey of northern NJ aquifers
• Begin assessing SW/GW interactions with an analysis of the Long Island data on the start of flows,' GW water-levels, and nitrate concentrations
• Plan and carry out the field reconnaissance of Long Island SW and ecology for potential site selection
• Develop and refine the GW transport model study-design strategy o Continue training and preparing for field activities, including outfitting 2 sampling vans
• And, of course, develop a written work plan for FY 1996.

Also, the discussion included various tasks analysing existing data and will likely result in reports. The following list of planned reports have a priority rating for a written report of * to *** or 1st to 3rd priority. Because the information is necessary to develop next year's work plan, we will try to finish all these analyses by early next year, but not all will end up in a report form by then. We hope to have reports labeled * completed and in review by the end of this calendar year.

* Lead-Paul Stackelberg, Occurrence and distribution of synthetic organic compounds on bed sediments of New Jersey streams and their relation to basin characteristics. (Journal article)

* Lead-Anne O'Brien, Distribution of trace elements on bed sediments of New Jersey streams and their relation to basin characteristics. (Journal article)

* Lead-Jonathan Kennen, Relation of benthic invertebrate and fish communities to water quality and basin characteristics of New Jersey streams. (Journal article)

** Lead-Bob Reiser, Factors affecting water and nutrient budgets of streams in New Jersey and Long Island. (USGS WRI report or journal)

* Lead-Rick Clawges, How do people affect groundwater quality? (A lay-reader style informational report summarizing NJ and LI GW studies)

*** Lead-Paul Stackelberg and Tom Barringer, Statistical evaluation of the NAWQA groundwater land-use survey approach for highly urbanized areas of New Jersey and Long Island. (Journal article)

We presented the results so far for the analysis of data on bed sediment chemistry and discussed how the existing biological data might be similarly analyzed. Good discussion ensued on these results, on equivalent efforts to assess existing benthic invertebrate and fish data, and on the need for results of all these analyses before any bed sediment and tissue survey can be planned, for that matter, before a work plan for most SW components can be finalized. Copies of Paul and Anne's overheads are included for your information and comment. The following bullets summarize conclusions and comments on questions for this session.

• Data were from the NJDEP-USGS cooperative water-quality network. Over 1000 bed sediment samples were collected since the late 1960's from about 300 sites in NJ. Lab analyses included solvent extractable chlorinated pesticides and PCBs, and acid extractable trace elements. Some analyses included total organic carbon.
• Organic compounds detected frequently were: DDT, DDE, DDD, chlordane, dieldrin, and PCBs. Detected infrequently and at low concentrations: aldrin, endosulfan, endrin, heptachlor, heptachlor epoxide, lindane, methoxychlor, mirex, perthane, and toxaphene. PCNs were not detected.
• The 6 organic compounds which are frequently detected tend to be widely distributed. Concentrations of organic compounds tend to be elevated in specific land use areas; for example, the highest median concentrations for each of the 6 most frequently detected organic compounds were in the most heavily urbanized area (Arthur Kill). Patterns of detection in similar land use areas varied; for instance, DDT (and metabolites) were detected much more frequently in the Coastal Plain agricultural areas than in northwestern NJ agricultural areas.
• Even though DDT was applied in urban and agricultural areas in the past, regression results indicate an inverse relation between agricultural land use and the detection of DDT above a 10 micrograms per kilogram threshhold? Perhaps this is related to how water gets to the stream (ie, quicker route in urban areas) or to the way DDT was used on agricultural lands? Does anyone have an idea why this might be?
• Although numerous explanatory variables produced regression models, the best fit logistic models related the presence of organic compounds to a combination of total population and one or two other land use catagories, such as urban, residential, and/or agricultural land use.
• Logistic regression indicated that the probability of detecting elevated trace element concentrations is related to upstream land use, population, geology (ie, physiography), and total flow from point sources.
• Elevated concentration relations to variable (by element):
• total population = Cu, Ni, Pb, Zn
• flow from point sources = Fe, Zn
• area in agricultural land = As, Mn
• area in New England Prov.(glaciated) = As, Cd, Cr, Pb, Zn
• negative with area in Coastal Plain = Fe, Zn
• NURE (National Uranium Resource Evaluation survey) data is an independent data set that may be useful to test regression models. Samples of bed sediment are largely from small basins and should reflect background levels. Caution: NURE used total digestion, Coop network used total extractable procedure, not equivalents. Karl Muessig's group has maps showing differences among physiographic provinces.
• Of the 7 basin types, should Big Timber/Cooper R. drainage area be considered as one? Although recognizably different now, the land use composition and population density are not all that different. Moreso, the composite drainage (and sites sampled) of the two represents a range of urban development in the Coastal Plain, so it fits its intended purpose.
• Do detection frequencies change over time? Perhaps a plot of multi- year detection frequencies will show if a trend exists or not. o How should we deal with the land use changes that took place since the GIRAS coverage was made (circa 1972)? Obtain a more recent coverage perhaps.
• Bed sediment chemistry data are highly variable. How much variation is due to sampling methodology, suggesting a more uniform approach in sampling is needed (if possible)? How much variation is due to changing sources and fluvial processes, suggesting more frequent sampling at fewer sites might help understand the differences over time?
• What would be gained by doing a typical NAWQA bed-sediment and tissue survey? We need to similarly examine the data on aquatic communities and compare results to the bed sediment analysis for multiple lines of evidence' in these relations and to determine if there is a corresponding impact of synthetic organics and/or elevated trace elements on community health. This will help establish if and/or where any further bed sediment and/or any tissue data should be collected.
• Can we design a study to determine the source of contaminants? New or old sources? Remobilization? Some focused reach studies, including storm sampling, would be needed. This will be considered in the NAWQA plan.
• What about issues on Long Island with build up in recharge basins? NAWQA should review existing work on this.

The afternoon session included initial discussions on what a groundwater study-unit survey of northern NJ might look like and on a proposal to use a model of flow and contaminant transport as a planning tool to test hypotheses/concepts and to help design data collection efforts for improved understanding of processes in this data rich study unit.

A preliminary plan for a GW flow and contaminant transport model analysis for LI and NJ was presented and discussed at length. The following are bullets from the session.

• How does this effort fit in the NAWQA design? We consider it a demonstration effort of a new approach to improve our understanding of GW contamination and will address the GW work group's statement -- "We know where the problems are, tell us what can we do about GW contamination?"
• General concepts and strategy for the model were presented by Art Baehr. The proposed modeling effort basically consists of two parts, development of a 3-D groundwater flow model (MODFLOW) of the area and then a multi-species water quality transport model which utilizes realistic aquifer flow patterns generated in the flow model.
• We are looking for 2 hydrologically discrete areas to model, one on LI and one in the NJ Coastal Plain. The spatial scale of the proposed investigations is about 100-200 square miles in size, a scale that provides good hydrologic boundaries and is regional in nature. The 2 areas receiving most discussion so far were (1) a highly developed and data rich area draining south of the LI GW divide in the vicinity or east of the Nassau/Suffolk County borders (thus providing an opportunity to test hypotheses related to an area with a strong historical perspective on GW quality management issues) and (2) a developing area with a history of intense agriculture along the Highway 55 corridor centered near Glassboro and Washington Twp., Gloucster County, NJ where additional development of the Cohansey Aquifer is under consideration (thus providing an opportunity to test hypotheses related to an area with strong water-supply planning/allocation issues). The surficial aquifers of the Metedeconk/Toms River area and the eastern Mercer/western Monmouth/western Middlesex Counties area were also mentioned. What of these or other areas might best fit everyone's needs? Do we really need 2 areas to model if a more conceptual approach to modeling the water quality is taken?
• We need to define the contaminants of most interest--nitrate, VOCs, pesticides, hydrocarbons from petroleum products were discussed. Emphasis seemed to be with nitrate and VOCs within the proposed study areas. Trace elements are beyond the scope of this study, but could conceptually be done with this model approach.
• The scope of this effort calls for the team to develop flow/contaminant transport models for the 2 areas and begin hypothesis testing over the next 12-18 months. That short time frame mandates the use of high quality data bases and use of existing hydrogeologic framework from previous flow modeling studies that exist for LI and the Coastal Plain of NJ.
• The model analyses also will be useful in designing other components of the NAWQA study including follow-up sampling of the study area `flow systems' to verify model constructs of major controlling processes. Data collection would address constituents of interest and other supporting data on age dating, isotopes, and microbial processes. In NAWQA jargon, these field efforts would be more areally focused than a typical land-use survey but more regional than a flow-path survey, in a sense, they are flow-system surveys of land use effects. Because this approach incorporates modeling as a planning step it should result in taking our understanding farther than sampling alone would.

A tentative date of December 7,1995 was chosen for the next liaison meeting. Enclosed are copies of several overheads used in the talks, fyi. Please write or call in any comments or suggestions (Mark at 609-771-3943, Anne and Paul at 3951, Art at 3978). Thank you.