MAJOR FINDINGS

Figure 5. Stream conditions for fish communities have improved from the 1970s to the 1990s at 41 of 88 sites in northern New Jersey.

Stream Conditions for Fish Have Improved Since the 1970s

Fish are useful indicators of environmental changes in drainage basins because they are sensitive to a wide variety of stresses including changes in water chemistry and flow, modifications in habitat and food, and landscape alterations resulting from urbanization and other human-related activities. Moreover, fish accumulate certain contaminants within their tissues over their entire life span (Fausch and others, 1990; Karr and others, 1987, Chang and others, 2000). One way to identify environmental changes in watersheds is to assess changes in stream condition by use of an Index of Biotic Integrity (IBI; blue text box below). On the basis of IBI results, conditions for fish in streams of the Delaware, Passaic, and Raritan River Basins have improved since the 1970s (fig. 5; table 2) (Chang and others, 2000).

Of the 88 sites assessed, stream condition improved at 46 percent, worsened at only 13 percent, and remained the same at 41 percent from the 1970s to the 1990s (fig. 5).

Many factors, especially the improved treatment of wastewater discharges during the 1980s and changes in land-use practices, may have contributed to the statistically significant increase in IBI scores in all three basins (table 2). Nonpoint sources were less of a factor than were point-source improvements; but as discussed on pages 7 and 8, urban nonpoint influences are still significant. The Delaware River Basin reflected the greatest improvement in stream condition, moving from a condition category of fair up to good (table 2).

A change in stream condition also can be assessed by comparing the percentage composition of the fish families (fig. 6). For example, the highly contaminant-tolerant white suckers made up a greater percentage of the fish communities in all basins during the 1970s than in the 1990s, especially in streams in the Delaware River Basin. In addition, during the 1990s, darters, of which many species (for example, the shield darter) are highly intolerant of human disturbance and habitat modification, increased in abundance in streams in the Passaic and Raritan River Basins and remained unchanged in streams in the Delaware River Basin.

Many minnow species are indicators of good stream condition because they rely directly on the availability of aquatic insects for survival. Thus, as streams become degraded and insects and their larvae become scarcer, the numbers of insect-eating minnows often decrease. Minnows made up greater than 42 percent of the community during the 1970s and 1990s in all three river basins (fig. 6). The percentage of minnows increased in Delaware River streams. Minnows decreased slightly in the Passaic and Raritan River streams, however, an indication that these two basins may still reflect some degree of impairment.

These differences, perhaps, may explain why the mean IBI score for streams in the Delaware River Basin increased slightly more than that for streams in the Passaic or Raritan River Basins (table 2). Other fishes (such as eels) were not included in this comparison because they are found over a large range in water-quality and habitat conditions, thus limiting their use as aquatic indicators (Kurtenbach, 1993). Chang and others (2000) provide information on collection methods, fish taxonomy, and variability between the 1970s and 1990s collections.

Figure 6. Changes in the percentage of fish families (that is, a decrease in suckers and a general increase in darters) captured in streams of the Delaware, Passaic, and Raritan River Basins during the 1970s (top) and the 1990s (bottom) indicate that stream conditions have improved.

Aquatic Invertebrate Communities Differ Naturally Across the Study Area

The New Jersey Department of Environmental Protection (1994b) established 43 reference or benchmark sites whose drainage basins have been minimally disturbed by human activity. A comparison of these sites indicates that the aquatic invertebrate communities in the southern New Jersey Coastal Plain are distinctly different from those in the north as a consequence of natural differences in environmental and physical conditions (table 3). Thus, natural variability in aquatic invertebrates needs to be taken into account when assessing communities across the entire study area. This finding helped focus our data-collection efforts toward gaining a better understanding of urban influences on northern New Jersey communities. In addition, analysis of LINJ NAWQA data indicates that aquatic-invertebrate communities on Long Island closely resemble New Jersey Coastal Plain communities because of similar physiographic and habitat conditions.

Fish and Aquatic Invertebrate Communities Are Impaired in Urban Areas

Figure 7. Aquatic-invertebrate-community data at more than 700 New Jersey AMNET sites (blue text box above) commonly indicate moderate to severe impairment in higher density urban areas (gray areas in fig. 1, p. 3).

Despite some overall improvements in stream condition (p. 5), fish and aquatic-invertebrate communities are commonly impaired in the urbanized parts of New Jersey (figs. 7 and 8; see also fig. 1 on p. 3 for a comparison with urban area). Community-impairment scores at more than 150 fish-sampling sites (Kurtenbach, 1993) and more than 700 aquatic-invertebrate sampling sites (for example, New Jersey Department of Environmental Protection, 1994a) in New Jersey were examined with respect to land use and other basin characteristics. The northwestern part of the State is less developed and consequently was least likely to reflect moderately or severely impaired aquatic communities (figs. 7 and 8).

The percentage of urban area in the drainage basin and the amount of treated wastewater flows upstream from sampling sites were the primary factors related to a severely impaired aquatic-invertebrate community (Kennen, 1999). Total amount of forest land in the basin, however, was a strong mitigating factor and increased the likelihood of an unimpaired aquatic-invertebrate community. Similarly, total population in proximity to the sampling site and the amount of urban land in a basin were found to be most highly related to poor fish-community conditions.

Figure 8. Fish-community data at more than 150 sampling sites in northern New Jersey indicate that streams in urban areas are in poor to fair condition (based on IBI scores; blue text box on p. 5).

The following section describes an extension of these analyses of available data through an integrated study of the specific factors that affect biological communities in urban environments.

Characteristics of Urban Landscapes That Affect Aquatic Communities

More than 400 landscape and environmental variables were aggregated for comparison with data describing fish, invertebrate, and algal communities at 36 streams in New Jersey and Long Island. Sites were chosen with drainage basins that ranged from 3 to 96 percent urban land. Analyses of these data indicated that increasing impairment of fish, aquatic-invertebrate, and algal communities was statistically related to components of the urban gradient (table 4; refer also to “Glossary,” p. 31).

Environmental factors such as annual peak discharge, amount of impervious road area, and population density were related to impairment in all three types of aquatic communities (negative in table 4). Some environmental factors such as point-source flow, urban growth during 1986–95, and impervious nonroad area were related to impairment in the aquatic-invertebrate community only (negative in table 4). In addition, total urban area in 1986 appeared to be important only for the fish community. Other studies have found that historical changes in land use may have significant implications for longer-lived organisms such as fish.

Figure 9. Unregulated impervious-area runoff directly affects water quality, habitat, and aquatic communities in streams and is exemplified by a storm-sewer pipe (left) that drains directly into the Saddle River at Ridgewood, N.J.

The presence of cobble substrate was a factor contributing to healthier fish and aquatic-invertebrate communities (positive in table 4). Degradation of cobble and other stream habitat in urban systems likely is related to increases in flow, channel erosion, and sedimentation common to minimally controlled urban stormwater runoff (fig. 9).

In fact, changes in hydrologic factors (such as decreases in base flow and increases in peak discharge and the flashiness of streamflow) play a major role in influencing the types and condition of aquatic communities present in a stream (table 4, fig. 10), in large part, by the way these changes in flow affect stream habitat. As a result, stream communities are continually stressed and rarely reach stable population levels in urban systems. Conversely, reductions in base flow resulting from changes in water-use and wastewater distribution practices greatly influence the suitability of a stream for many types of organisms (Klein 1979). The area of forest and wetlands in the drainage basin was a positive factor in the health of aquatic-invertebrate communities (tables 4 and 5). Forests and wetlands play a major role in maintaining a healthy supply of water, food, and habitat for disturbance-intolerant and highly desired species. Thus, forest and wetlands are able to help mitigate the undesirable effects of other human-induced landscape alterations.

Figure 10. Reductions in base flow of streams resulting from changes in up stream land-use and water-use practices affect the distribution of aquatic species in streams. For example, the difference in wetted habitat for median-flow conditions (upper photograph) and low-flow conditions (bottom photograph) at Neshanic River at Reaville, N.J., is substantial, although not entirely related to human activities.

Organochlorine Compounds Were Detected in Streambed Sediment and in Fish Tissue

Even though the use of many organochlorine compounds has been discontinued, the widespread historical application and the environmental persistence of compounds such as chlordane, dieldrin, DDT, and PCBs have led to frequent detection in streambed sediment and in whole-fish tissue samples in Long Island and New Jersey (Stackelberg, 1997; Long and others, 2000). Analyses of available bed-sediment chemical data for nearly 300 sites (Stackelberg, 1997) indicate that chlordane and dieldrin concentrations in bed sediments were highest in urban areas, reflecting their past use for termite control. Concentrations of DDT and PCBs were highest in urban and industrial areas, reflecting their principal historical use as an insecticide and an industrial chemical (in hydraulic lubricants and heat-resistant oils in electrical transformers), respectively. No significant relations were found, however, between chlorinated hydrocarbon concentrations in bed-sediment and whole-fish samples (white sucker, the chosen target species for this analysis) or between those in whole-fish samples and land use (Long and others, 2000).

Table 5. Comparison of biological status at 36 Long Island–New Jersey stream sites to 140 NAWQA sites nationwide indicates that the high scores relative to other NAWQA sites are associated with urban watersheds. Sites are ordered from top to bottom in descending percentage of urban land (urban gradient)

Of the eight streams sampled in fall 1997, one-half of all bed-sediment samples contained concentrations of chlordane, DDT, and PCBs that exceeded threshold effect levels (fig. 11; blue text box on p. 12; Long and others, 2000). Concentrations of chlordane and PCBs in whole-fish samples from many of the sites exceeded established guidelines for fish-eating wildlife. Some of the detected compounds are known to pose human health risks; however, concentrations in whole-fish samples from this study are not directly comparable to concentrations in edible portions (fillets) that are used to establish U.S. Food and Drug Administration (FDA) action levels for human consumption (U.S. Food and Drug Administration, 1992). Nonetheless, the FDA action levels for human consumption of total chlordane [300 µg/kg (micrograms per kilogram)], total DDT (5,000 µg/kg), and total PCBs (2,000 µg/kg) in edible portions of fish were not exceeded in the whole-body samples at any of the sites. Organochlorine compounds in lake-sediment cores and in water samples are discussed on pages 14 and 20, respectively.

Comparisons of invertebrate and algal status at 36 LINJ sites to their status at 140 selected NAWQA sites nationwide were made using nationally derived indicators (described in blue text box below). These indicators were selected because of their ability to discriminate, in a predictable way, human influences on the environment. These indicators also have the desirable characteristic of sensitivity to environmental stressors and a low dependency on natural variability such as elevation, stream size, and ecoregion. The results discussed in this section are for the sole purpose of a national comparison and refer only to a population of NAWQA Program sampling sites. They are not designed to be used as benchmarks for other State, regional or national studies because the aquatic community indicators have not been calibrated for such purposes. Biological indicators of water quality in a national context—The selected biological indicators respond to changes in stream degradation. Degradation can result from a variety of factors that modify habitat or other environmental features such as land use, water chemistry, and streamflow. Algal status focuses on changes in the percentage of certain algae in response to increasing siltation and often appears to correlate closely with increasing nutrient concentrations. Invertebrate status is the average of 11 invertebrate biometrics that summarize changes in richness, tolerance, trophic conditions, and dominance associated with water-quality degradation. Fish status, which sums the scores of four fish metrics (percent tolerant, omnivorous, non-native individuals, and percent individuals with external anomalies), was not used for this national comparison because values for the full suite of sites were not available. For all these indicators, higher values indicate a more degraded system. About one-half of LINJ status scores ranked in the middle one-third (yellow squares, table 5) of the 140 NAWQA sites sampled during 1996–98. Some LINJ status scores (orange squares), however, were in the upper one-third, and a few sampling sites such as the Elizabeth River at Hillside, Rahway River at Washington Park, and Passaic River near Chatham, N.J., were among the highest scores nationally for algae and invertebrates. All these sites fell in the upper part of the urban gradient, or in basins with greater than 47 percent urban land use. These sites were dominated by disturbance-tolerant aquatic invertebrates such as worms and midges and supported few if any disturbance-sensitive invertebrates such as mayflies, stoneflies, and caddisflies. In addition, highly silt-tolerant algae made up from 86 to 98 percent of the overall community abundance. The presence of these tolerant forms reflect significant levels of disturbance and highly degraded instream and riparian habitat. About 35 percent of the LINJ status scores were among the lowest one-third nationally (blue squares, table 5). The Rockaway River at Boonton (figure top right), Lamington River near Pottersville, South Branch Raritan River at Arch Street, Spruce Run at Glen Gardner, Pequannock River at Riverdale, and Flat Brook near Flatbrookville, N.J., had some of the lowest scores nationally for algae and invertebrates (table 5). Land use in the basins of these sites is less than 34 percent urban. These sites were among the most diverse sampled in the LINJ study; 50 to 79 percent of the invertebrate community was composed of intolerant organisms such as mayflies, stoneflies, and caddisflies. In addition, silt-tolerant algae made up less than 19 percent of the algal community abundance at all sites. Sampling aquatic invertebrates in the Rockaway River at Boonton, N.J. Two sites, Beden Brook near Rocky Hill and Doctors Creek at Allentown, N.J., were among the sites with the highest status scores nationally (most degraded) for algae and invertebrates, respectively (table 5), yet corresponding basins had relatively little urban land (less than 25 percent). In addition to nonpoint-source influence, these sites likely reflect additional degradation resulting from wastewater-treatment-plant effluent. Doctors Creek also has a high proportion of agricultural land and far less forest than many of the sites on the lower end of the urban gradient. Agricultural land is known to produce high levels of sedimentation in surface water and high levels of nutrients in surface water and ground water (see section on nitrate concentrations, p. 15), which have historically been linked to aquatic-community degradation (Culp and others, 1986).

Trace-Element Concentrations Were Elevated in Streambed Sediment, Fish Tissue, and Ground Water

Human activities have accelerated the release of trace elements to aquatic environments through point- and nonpoint-source contamination (blue text box below). Analyses of available bed-sediment chemical data for nearly 300 sites (O'Brien, 1997) indicated that trace-element concentrations in bed sediments were generally higher in the northern New Jersey physiographic provinces (related to urban development and geologic availability) and lower in the Coastal Plain (related to lower organic content of sediments, geologic availability, and stream-water pH). Higher copper, lead, and zinc concentrations were correlated with increased population density; zinc was correlated with wastewater-treatment-plant flows; higher arsenic was correlated with increased agricultural land use; and higher chromium was correlated with certain geologic deposits (especially in the New England Physiographic Province).

All eight of the streams sampled during fall 1997 were found to have at least two trace elements that exceeded TELs in bed sediment (fig. 11; Long and others, 2000; blue text box on TELs and PELs below). Concentrations of eight trace elements at Rahway River near Springfield (fig. 11) exceeded TELs; arsenic, cadmium, and copper also exceeded PELs.

Figure 11. Concentrations of selected trace elements and organochlorine compounds in bed sediment commonly exceeded TEL and PEL guidelines, especially at urban sites. (Background concentrations were estimated using method of Velz (1984); see blue text box below concerning TELs and PELs.)

Copper, manganese, mercury, and selenium were detected in fish livers at all sites. No significant relations were found between trace-element concentrations in fish livers and those in bed sediment or between concentrations in fish livers and any land-use category or population density (Long and others, 2000).

Arsenic was detected in 17 of 22 samples collected from domestic-supply (household) wells completed in fractured bedrock aquifers of the Piedmont Physiographic Province, where geologic formations are known to contain arsenic-bearing minerals. Concentrations ranged from 1 to 57 µg/L. The current drinking-water standard for arsenic is 50 µg/L, but the USEPA has proposed a tenfold decrease in this standard to 5 µg/L. Six of the samples (27 percent) contained arsenic in concentrations equal to or greater than 5 µg/L.

Concentrations of radium (sum of Ra-226 and Ra-228) above the drinking-water standard were found in 33 percent of 170 wells sampled in the surficial aquifer system in southern New Jersey (Szabo and others, 1997). The highest radium levels were in areas where acidic waters are associated with surficial sediments that contain radium from geologic sources (primarily the Bridgeton Formation) and agricultural areas where use of nitrogen fertilizers and lime is heavy. The leaching and nitrification of applied fertilizers, which increase the dissolved-solids content and acidity of ground water, are likely mechanisms by which radium is mobilized from surficial sediments and transported to ground water. More recently, the short-lived isotope radium-224 was detected in samples from this surficial aquifer system in waters affected by the same chemical processes (Zoltan Szabo, U.S. Geological Survey, written commun., 2000).

In a 59-well subset of the 170 wells above, mercury was detected at concentrations above the drinking-water standard of 2 µg/L in about 10 percent of the samples. The source is suspected to be linked to past use of mercury-based pesticides. Mercury may have increased mobility as a result of the natural acidity of waters from the surficial aquifer in southern New Jersey, as well as widespread contamination of these waters with chloride from road salt, septic systems, and other discharges; but the mechanisms are as yet not well understood.

Analyses of lake sediment cores are an effective method for evaluating water-quality trends. (See blue text box below.) Cores were extracted from three lakes in northern New Jersey and one on Long Island as part of a national lake-coring study (Callender and Van Metre, 1997). Sediment was dated by use of cesium-137, a by-product of nuclear-weapons testing. Trace elements (arsenic, cadmium, chromium, lead, mercury, nickel, and zinc) were detected throughout the cores of all four lakes, with concentrations of most elements elevated in the three urbanized basins relative to the largely forested Clyde Potts Reservoir basin. Until concentrations of lead peaked in the 1970s, lead and zinc concentrations were highly correlated over time with the population in the vicinity of the lake. Population is an indicator of vehicular traffic, which in turn is an indicator of gasoline use (a probable lead source) and tire wear (a probable zinc source). Removal of lead from gasoline by the Clean Air Act resulted in a general decrease in sedimentary concentrations of lead since the mid-1970s phaseout (fig. 12); however, lead concentrations remain elevated compared to detections at the base of the cores. Zinc concentrations are generally increasing in the three urban watersheds in response to increasing population and traffic density, but not in Clyde Potts Reservoir, which is least affected by traffic. Detectable concentrations of chlordane, total DDT, total polycyclic aromatic hydrocarbons (PAHs) and total PCBs were found in all lake cores; dieldrin was detected in only two lake cores. Generally, concentrations of organochlorine compounds began to decline after regulatory action discontinued organochlorine production and use in the 1970s and 1980s, but the persistence of these compounds may mean that a substantial amount of time must pass before they are purged from the basin. Concentrations of PAHs, however, are elevated and generally increasing over time in sediment cores (fig. 13), presumably as a result of increased vehicular and other fossil-fuel use associated with urban development. Concentrations of PAHs were lowest in Clyde Potts Reservoir (least traffic influence). Of particular note are the low concentrations of PAHs in Packanack Lake sediments in the 1930s, when automobiles were comparatively rare and before the watershed was urbanized. These trends have been observed in other urban lake-sediment cores across the United States (Callender and Rice, 2000). Figure 12. The mid-1970s phaseout of lead from gasoline as a result of the Clean Air Act has resulted in decreased basin inputs of lead concentrations in lake sediment.     Figure 13. Concentrations of total polycyclic aromatic hydrocarbons (PAHs) are elevated and generally increasing in sediment cores as a result of increased vehicular traffic and fossil-fuel use.