Kimberly Yandora

City of Greensboro, Storm Water Services, 401 Patton Avenue, Greensboro, NC 27406

Rapid bioassessment of macroinvertebrates was conducted at thirty-one sites within the urban watershed of Greensboro, North Carolina during the summer of 1997. Assessment at each site included physicochemical parameters, habitat score, and the following indices: total taxa richness, North Carolina biotic index value (NCBI), EPT (Ephemeroptera, Plecoptera, and Trichoptera) abundance, EPT richness, ratio of EPT and Chironomidae, percent Tubificidae, and percent dominant species. Sites up-stream of urban activity showed high diversity and richness of aquatic communities and overall good water quality. Poor to fair water quality ratings were seen downstream of urban activity. However, the condition of biotic communities was directly related to habitat and water chemistry. Habitat is degraded in urban areas due to dredging, channelization, and impaired riparian buffer zones that contribute to poor species diversity. The results of the bioassessment monitoring program lead us to the conclusion that physical and chemical data of storm events and baseflow stream conditions cannot fully assess the effects of urbanization on small order streams.

We have been monitoring storm water runoff for four years and ambient in-stream conditions for two years to establish water quality history. Storm water runoff from developed land showed elevated levels of pollution whereas ambient in-stream conditions showed much lower levels. Using benthic macroinvertebrates as indicators of localized conditions aids interpreting water quality data because they lead stationary lives and respond quickly to stress from storm events and illegal dumpings. This study illustrates the importance of biological data in conjunction with physicochemical data to assess water quality and to characterize impacts from urban runoff.

Greensboro, North Carolina is located at the headwaters of the Cape Fear River Basin in the Piedmont ecoregion and has 509 linear miles of streams (NCDEHNR 1995). Greensboro is a rapidly growing city with a population of over 200,000 and an area of 109 square miles. The City of Greensboro was issued a National Pollutant Discharge Elimination System (NPDES) Permit in 1994. This permit requires the monitoring of storm water runoff within the city to characterize pollutant loading from different land uses and to estimate annual pollutant loading. In addition, the permit focuses on the elimination of non-point source pollutants through identification of every outfall in the city with a focus on industrial areas. The City of Greensboro also conducted monitoring of ambient in-stream conditions to establish baseline water quality and conducted acute toxicity testing on storm water. However, the effects of spills, illegal discharges and other episodic events cannot be qualified or quantified by any of these methods.

Data on runoff from storm events indicated high levels of pollutants entering into the small streams of Greensboro, but baseline in-stream data showed relatively low amounts of chemical pollutants. It was determined that benthic macroinvertebrate would be good indicators of localized conditions. In addition, many sites throughout the city could be studied with relative ease. Macroinvertebrates integrate the short-term environmental variations because they respond quickly to stress from spills and storm events and lead stationary lives. Macroinvertebrates, rather than fish, were selected because many of the first and second order streams in Greensboro might not be able to support fish assemblages whereas invertebrates are present in most streams (EPA 1996).

We sampled benthic macroinvertebrates at 31 sites within and draining into the watershed of the City of Greensboro. North Buffalo Creek and South Buffalo Creek are located in the heart of the oldest and most urbanized areas of the city with the highest amount of development and impervious area. Streams in these sections often have been channelized and dredged in the past. South Buffalo Creek has the most recent construction activity and as a result more sediment loading. Reedy Fork Creek is located in the north reaches of the city limits and drains primarily undeveloped or agricultural areas. A series of five reservoirs have been built on this stream to provide drinking water to Greensboro. In the last ten years, urban sprawl has started spreading into these basins with increased residential, commercial and industrial activities. In addition, the East Fork Deep River and Bull Run are within Greensboro City limits and discharge to the City of High Point’s water supply.

Emphasis was placed on the major tributaries supplying Greensboro’s and High Point’s drinking water supplies. Efforts were made to spatially distribute sites over the entire area within the city limits of Greensboro. Macroinvertebrate samples were collected on the North and South Buffalo Creeks at sites that are monitored for storm water runoff and in-stream baseflow conditions. In addition, tributaries were selected to provide basinwide coverage. Very little monitoring had been conducted on the Reedy Fork Creek. All the main tributaries to the reservoirs and downstream of the reservoirs were sampled. In addition, one site on East Fork Deep River and Bull Run were sampled to estimate the condition of these streams as they leave the city limits of Greensboro.

Three sites were selected as reference sites in relatively undeveloped locations: Bryan Park, Battleground, and McKnight Mill. Each is a first, second, and third order stream, respectively. These sites are important since replicate sampling was not conducted and will be used for comparison.

Benthic sampling followed a modified version of the EPA Rapid Bioassessment Protocol II using single habitat approach with 1-meter kick net with 500-m m mesh openings (EPA 1996). A 100-m reach representative of the stream was selected. All samples were collected from the riffle zones of streams in areas where there was the best canopy coverage and side bank vegetation to portray the best overall sample results. Whenever possible, the site was at least 100 meters upstream of roads or bridge crossings and had no major tributaries discharging to the site. Two or three kicks were sampled at various velocities within in the stream reach. Large rocks and logs in the area where dislodged and washed off within the net. From the net, the sample was placed into a 500-m m opening sieve bucket where leaves, twigs and other large debris were washed off and discarded. The remaining debris and sample was placed in plastic containers and preserved in 90% ethanol. All organisms were sorted from debris in the laboratory and then 100-organism sub-sample was randomly selected from a standardized grid. The one-hundred organism sub-sample was properly labeled and preserved in glass containers in 90 % ethanol. Three of the sites were samples twice to provide quality assurance.

Sorted samples were sent to a qualified contracted laboratory where organisms were identified to the lowest practical taxon, usually species. Specimens too immature or damaged to identify below the level of genus were reported to the lowest known level. Identifications were checked by having 10% of the samples randomly selected and identified by another biologist. Tolerance values and functional feeding groups were reported. Hilsenhoff tolerance values were used when North Carolina’s tolerance values were not available. North Carolina’s tolerance values range from 0 for organisms very intolerant of organic wastes to 10 for organisms very tolerant to organic wastes.

The following statistics were calculated for each site: total taxa richness, North Carolina biotic index value (NCBI), EPT Abundance, EPT Richness, Ratio of EPT and Chironomidae, Percent Chironomidae, Percent Tubificidae, and Percent dominant species (EPA 1996, NCDEHNR 1997, MCDEP 1996).

Taxa richness is the simplest measure of diversity. The total number of species collected in the 100-organism subsample was recorded to measure taxa richness. Taxa richness decreases with a decrease in water quality as the less tolerant species are eliminated. Bioclassification criteria developed by North Carolina Department of Environment, Health, and Natural Resources (NCDEHNR) for the North Carolina Piedmont for the standard qualitative sampling method is listed in Table 1. This bioclassification criterion is based on values from summer collection (June - September). These ratings reflect effects of chemical pollution but poorly assess the effects of sediment pollution.

NCDEHNR developed the NCBI which accounts for differences in stream size, seasonal variations and ecoregions to complement taxa richness (NCDEHNR 1997, Lenat 1993). The NCBI is intended to examine the general level of pollution, regardless of source.

The NCBI is derived using the following formula:

where Tv_i is the tolerance value of the ith taxa, N_i is the abundance of the ith taxa (1,3 or 10) and N is the sum of the abundance values. The abundance information for each taxon is tabulated at either RARE (1-2 specimens), COMMON (3-9 specimens) or ABUNDANT (> 10 specimens) and given the value of 1, 3, or 10 respectively. The bioclassification criteria developed by NCDEHNR for the NCBI (after seasonal corrections) for the North Carolina Piedmont are listed in Table 2 (Lenat 1993).

Good biotic conditions would be reflected in communities with an even distribution among all four major groups. Skewed populations having a disproportionate number of Chironomidae relative to the more sensitive organisms (Ephemeroptera, Plecoptera, and Trichoptera) indicate environmental stress (EPA 1989).

The percentage of the family Chironomidae in the sample represents whether a stream is oligotrophic or eutrophic. A sample in which greater than 50% is Chironomidae suggests eutrophic conditions. Some species of Chironomidae are also tolerant to heavy metals. Percentage of Chironomidae will increase with a decrease in water quality.

An abnormally high percentage of Tubificidae accompanied by abnormally low values for percent Chironomidae indicates toxicity form urban runoff or insecticides that are toxic to arthropods. High tubified percentages accompanied by large Chironomidae populations indicate a serious organic problem.

A community dominated by relatively few species would indicate environmental stress. Dominant species greater than 35% indicates poor water quality, between 23%-35% indicates fair water quality and less than 25% indicates good water quality (EPA 1996).

Water samples were collected in plastic containers from locations at the middle of stream prior to macroinvertebrate sampling. Samples were preserved on ice and analyzed within 24 hours for nitrate-nitrogen and reactive phosphorus using the Hach Portable DR2000 Spectrophotometer. Turbidity, conductivity, pH, temperature and dissolved oxygen where measured at each site. Upon completion of sampling, a habitat assessment of each site was conducted using a format developed by EPA Rapid Bioassessment Protocol (EPA 1996). A numerical habitat score was calculated for each site. Habitat assessments were summed to obtain overall habitat score: optimal (260-201), sub-optimal (200-136), marginal (135-71), and poor (<70). Stream order for each site was determined using USGS topographic maps.

Greensboro, NC annually receives over 38 inches of rain. Input of pollutants from storm water runoff is frequent and a source of pollutant loading in our streams and waterways. This was evident in the results from four years of land use storm water quality data that showed heavy metals, fecal bacteria, and solids were the greatest impacts from urban runoff. These parameters frequently exceeded NCDEHNR action limits and standards for in-stream concentrations. However, acute toxicity of first flush samples using the fat head minnow (Pimephales promelas) showed no mortality. To complement this information, two years ago monthly sampling of ambient stream conditions was started to determine baseflow conditions. These samples were taken on the second Tuesday of each month to reflect any weather conditions but most likely reflected dry weather conditions in the stream. Fecal coliform levels continued to be elevated during baseflow conditions. Aluminum and iron, which were not tested in storm water runoff, were at levels above the NCDEHNR’s criteria, but were attributed to local soil types. The heavy metals, copper, lead, and zinc, which were prevalent in storm runoff, were much lower in ambient conditions. It is hypothesized that these particles quickly settle out of the water column into the sediment. As expected, solids were much lower in ambient stream conditions as was biochemical and chemical oxygen demand (BOD and COD, respectively).

Selected results and the associated metrics are listed in Table 3. The purpose of biological sampling was to sample citywide and to help determine the areas where more in-depth monitoring could be conducted or sites where future development may adversely affect stream conditions.

At each site where macroinvertebrates were sampled, nitrate-nitrogen and reactive phosphorus, pH, conductivity, turbidity, dissolved oxygen, and temperature was taken. Results were consistent with ambient in-stream data. Nutrient levels were found at very low concentrations. Phosphorus ranged between 0.0 mg/l to 0.75 mg/l and nitrate ranged from 0.2 mg/l and 1.5 mg/l. Temperatures were slightly elevated at sites with little canopy cover in comparison with site that had vegetative cover. Dissolved oxygen (DO) was not below 4 mg/l at any the sites. The lowest DO was 4.99 mg/l and as high as 9.73 mg/l. Turbidity was below 25 NTU at all sites except three sites that ranged between 90 to 102 NTU. These sites are associated with construction activities. Conductivity generally ranged from 64 m mhos/cm to 506 m mhos/cm and one site with 996 m mhos/cm. Sites with high conductivity were observed to be associated with industrial activities.

The purpose of the storm water monitoring program is to assess the overall health of the streams in Greensboro. NPDES permit requires monitoring of storm water runoff for the purpose of land use characterization and pollutant load estimates. However, this does not adequately characterize the health or condition of city streams. It does not take into account stream habitat or biological communities. This monitoring provides instantaneous chemical and physical data but does not indicate long term or continuous effects. Biological communities are directly affected by these parameters in addition to upstream and downstream activities such as piping of streams, dams, impoundments, construction activities, and stream crossings.

Great variation in habitat was seen throughout the city. For this study, the best overall habitat area was sampled in the different stream reaches. Habitat scores were lower at sites that received good-fair and fair biotic ratings in comparison to sites rated good and excellent. Some sites had good riparian buffers and canopy cover but lacked adequate substrate and bank stability. Historic practice in the urban setting was to channelize and dredge city streams to convey the water as quickly as possible out of the city to minimize flooding. In addition, riparian zones were maintained mowed lawns and bank vegetation was scarce. These practices have led to the destruction of biological communities. Current City policy is to restore vegetative riparian zones and to stop dredging stream channels. Unfortunately, we are still left with the damage from the past. Now the latest problem seems to be the result of construction and development activities that continues to increase sediment and flow to the streams.

Endless number of metrics and indices exist to analyze macroinvertebrate samples (Resh et al 1995, Thorne and Willliams 1997, Washington 1984). We have chosen to follow the procedures outlined by NCDEHNR (1995). In particular, bioclassifications have been based on the NCBI to support water quality assessment. Other metrics were used to help interpret the overall quality of the site. Of all the metrics calculated, the most useful were NCBI, taxa richness, EPT abundance, and percent Chironomidae.

EPT Richness showed very little difference between individual site with NCBI ratings good, good-fair, and fair. However, impaired and poor sites had EPT richness and abundance values of 1 and 0 indicating absence of mayflies, stoneflies and caddisflies. However, this absence was already noted in the NCBI and taxa richness. Therefore, we would not recommend using these metrics alone.

Tubificidae populations were rare, only being found at seven sites comprising less than 5% of the community. Since Tubificidae were not present at many of the sites this metric is not useful. Similarly, percent dominant species and ratio of EPT to Chironomidae did not produce distinct results. The greatest tubificidae population was 26% of the community at Caesar, which had poor ratings from all metrics. These results indicated the site suffers from sever organic pollution and eutrophic conditions.

South Buffalo Creek exemplifies the degradation of macroinvertebrate communities along the stream continuum. Big Tree is a second order stream location on South Buffalo Creek with an excellent NCBI rating influenced by residential land use. Its habitat score was rated sub-optimal with no channel alterations. Boston Road is slightly downstream where South Buffalo Creek is a third order stream. This site had a NCBI rating of good even though it is located downstream of two heavy construction areas with high amount of sediment loading. Hillsdale is further downstream on South Buffalo Creek, which receives runoff from an older commercial area with shopping mall, restaurants and office parks. The stream channel has been dredged and the buffer zone and banks are mowed regularly. The NCBI value at the Hillsdale site was only rated fair. Slight improvement was seen downstream at the Trestle site where the NCBI rating was good-fair. Most likely the reason the NCBI rating was improved was better habitat conditions and changes in land use. However further downstream on South Buffalo Creek at the McConnell site, the location was rated fair with impaired substrate and quality even though the surrounding land use was not developed. This site suffers from the affects of upstream urban activities.

Sites along North Buffalo Creek and its tributaries also showed degradation along the stream continuum. The Arboretum site is the farthest site upstream on the main channel. This site location is a third order stream where the NCBI rating was good-fair with habitat rating of sub-optimal. The stream degraded slightly at Lake Daniel where the NCBI rating was only fair and the habitat score was reduced. Surprisingly, the NCBI rating upstream of the City’s wastewater treatment plant (WWTP) was rated good-fair. Previous investigation by the NCDEHNR reported a poor NCBI rating. Water quality data showed dissolved oxygen of 4.99 mg/l and conductivity of 996 m mhos/cm. Two tributaries draining into North Buffalo Creek at Caesar Park and White Street indicated poor NCBI ratings. Both are affected by industrial as well as commercial runoff.

Sites located in the various tributaries to the water supply lakes were rated excellent, good and good-fair by NCBI. EPT species were well represented and Chironomidae percentages were low at all sampling locations. Development was restricted in these areas requiring best management practices on new development. However, development has steadily increased. This preliminary data on benthic communities serves as a baseline for change as development continues and the watershed changes.

Sediment loading occurs from construction activities especially during storm events. Benthic macroinvertebrates are able to withstand short-term increases in suspended sediments; however, continuous high levels of sediment may have adverse effects. The sediment effects macroinvertebrates by changing substrate, causing respiration difficulties, lowering oxygen concentrations and reducing food value. Chironomidae may increase because they use fine sediments in the construction of cases and tubes (Wood and Armitage 1997). Therefore, higher percentage of Chironomidae would be expected in South Buffalo downstream of construction activities. However, Boston, Hillsdale, and Trestle, which had high turbidity, had relatively low Chironomidae percentages. Sediment was evident at most sites even in the upper reaches of the water supply watershed that had almost no cobble substrate. Sediment in these areas was from some construction and bank erosion. The effect of sediment and erosion of biota still needs to be studied.

It is not surprising that urbanization causes degradation of water quality, habitat and biotic communities in streams. However, we have concluded that historic water chemistry monitoring does not provide enough information to assess completely the condition of aquatic ecosystems. The rapid bioassessment protocol indicated if water quality, substrate, riparian buffer, or channel alterations have impacted the site. Sites thought to be severely degraded from the appearance and perceived water quality actually indicated good-fair biotic communities. Conversely, sites in the water supply watershed of the City thought to be relatively pristine having good water quality showed lower water quality, habitat and biotic community diversity than expected. Macroinvertebrates are an important monitoring tool to measure continuous and chronic effects from pollution, stream degradation from storm water runoff and point source discharges, and indicators of stream recovery. Data on invertebrate communities in conjunction with habitat and water chemistry data will provide the necessary tools for monitoring impacts to streams and other aquatic systems.

*Reference Sites