Lawrence B. Cahoon, Professor

Department of Biological Sciences

University of North Carolina at Wilmington, Wilmington, NC 28403

Phone: (910) 962-3706; Fax: (910) 962-4066; E-mail: cahoon@uncwil.edu

Janice E. Nearhoof, Research Associate

Department of Biological Sciences

University of North Carolina at Wilmington, Wilmington, NC 28403

Water quality monitoring in a 55-square mile coastal region in southeast North Carolina is providing information about water quality and pollution sources prior to construction of a regional wastewater treatment system with an integral stormwater management program. Monitoring data provide a statistical basis for evaluating changes in water quality as the regional system is constructed and as development in this rapidly growing area proceeds. The effectiveness of pollution control approaches, including stormwater management measures, elimination of septic tanks for waste treatment, and more innovative measures such as land application of treated wastes will be evaluated by use of the data base and the statistical description it provides of each monitoring location and drainage unit within the region.

Fecal coliform bacteria, total nitrogen, total phosphorus, chlorophyll a, total suspended solids, turbidity, dissolved oxygen, percent saturation, pH, temperature, and salinity have been measured at 36 monitoring locations, which cover several distinct drainage units, some of them heavily developed and others relatively undisturbed. Variation in the characteristics of the locations, including land uses, human densities, and drainage pattern, results in few significant correlations between water quality parameters over the whole region that might otherwise be expected. Several parameters display marked seasonality as well. Processes and events in the immediate vicinity of each monitoring location clearly drive significant excursions from statistically average conditions, although stormwater runoff events have yielded a very noisy signal at best. However, careful data comparisons have already enabled identification of likely sources of water quality problems. Consequently, predictions of water quality and evaluations of impacts are very location-specific, but promise to be quite useful as the data base expands and additional information is considered.

The South Brunswick Water and Sewer Authority (SBWSA) was created in 1993 as a regional governmental entity by an interlocal agreement among the incorporated towns of Sunset Beach and Calabash and Brunswick County for the purpose of developing a regional sewage treatment system. Rapid development along North Carolina’s coastline is driving the need for regional wastewater treatment and disposal systems as conventional septic systems become problematic at higher densities, especially in many coastal soils, and incorporated coastal municipalities lack the revenue base to build centralized waste treatment systems.

Rapid development raises other environmental and social concerns, however, and opposition to regional wastewater systems has developed around two arguments. First, centralized sewage treatment permits development of areas with soil types inappropriate for septic tanks, permitting higher population densities, which by themselves drive various secondary socioeconomic impacts, such as the need for increased and improved infrastructure and for revenue sources to support these needs. Second, experience in coastal North Carolina and elsewhere has demonstrated that centralized wastewater treatment systems, by driving increasing population density, also have the indirect effect of increasing the area and proportion of hardened surfaces, causing increases in stormwater runoff and resulting nonpoint pollutant loadings to surface waters (Mallin et al., 1998).

Both arguments about negative secondary impacts of centralized wastewater treatment have been invoked by opponents of the SBWSA regional wastewater treatment plan. Consequently, SBWSA has agreed to abide by such population density restrictions as the State of North Carolina’s Department of Environment and Natural Resources may include in its permitting process and has incorporated a Stormwater Management Program in its overall project plan. The Stormwater Management Program includes a water quality monitoring program, which was begun in 1996 through a contract with the University of North Carolina at Wilmington (UNCW).

The aims of the water quality monitoring program are to:

This paper describes some of the results of the water quality monitoring program, the uses to which those results have been put prior to construction of the SBWSA treatment systems, and some of the broader lessons offered by patterns in the data.

Water quality monitoring focuses on surface waters throughout the SBWSA 201 Facilities Planning Area, a 55 square mile portion of coastal Brunswick County near the North Carolina-South Carolina border that encompasses estuarine waters (including a portion of the Atlantic IntraCoastal Waterway (AICW), tidal and freshwater wetlands, a variety of natural and artificial ponds and lakes, and portions of several coastal river basins (the Caw Caw River and the Shallotte River), as well as the entire Calabash River basin (Fig. 1). Twenty two monitoring locations were established in October, 1996, nine more were identified and sampled beginning in February, 1997, and an additional five were added in September, 1997. Each monitoring location was chosen to represent water quality conditions in a particular water body or in a portion of a drainage. The criteria for selecting each location were that each had to be representative of its area or drainage, be safely and legally accessible, and have standing water year round. Sampling at three monitoring locations has been discontinued after determining that they failed to meet one or more of these criteria. Owing to the need to have data about stormwater runoff effects on water quality, our sampling program uses a routine schedule that over time has allowed us to sample during both storm event and non-event periods.

Eleven water quality parameters are measured at each of the monitoring locations, which are visited in groups of eight or nine once or twice per week so that every monitoring location is sampled about every three weeks. Dissolved oxygen (mg/liter), percent oxygen saturation, pH, temperature (^oC), and salinity (ppt) are measured in situ simultaneously with a YSI Model 85 meter. Turbidity (NTU) is measured on a single bottle sample with an HF Scientific DRT-15CE turbidometer. Total suspended solids (mg/liter) are measured on triplicate water samples by gravimetry (APHA, 1995). Chlorophyll a (m g/liter) is measured in triplicate with a Turner 10-AU fluorometer following Welschmeyer (1994). Fecal coliform bacteria (CFU/100 ml) are measured by membrane filtration (MFC) using the standard multiple dilution method (APHA, 1995). Total nitrogen (m g/liter) and total phosphorus (m g/liter) are measured on triplicate water samples by the simultaneous persulfate digestion method of Valderrama (1981) followed by colorimetric analysis on an Alpkem Flow Solution 3000 autoanalyzer. All sample collections are logged in field notebooks following standard QA/QC procedures and returned to the laboratory using proper chain of custody procedures. Field sampling personnel use cameras to document sampling procedures, sample collections, and field conditions at the monitoring locations. All laboratory analysis data are logged immediately in lab notebooks, recorded in dedicated computer files, and reported to SBWSA by fax or in monthly data reports. Unusual data indicating problems are reported immediately. Data, methods descriptions, and data interpretations are also posted on a World Wide Web page created to inform the public about this project: (http://www.uncwil.edu/cmsr/
waterQ).

We initially expected that some of the water quality parameters measured at our monitoring locations would be correlated or at least show some relationship. For example, we expected that two different measures of the material suspended in water, turbidity and total suspended solids, would closely correlate. However, simple plots of the values of these parameters measured simultaneously and pooled for all monitoring locations revealed no clear relationship (Fig. 2). This lack of a relationship becomes more reasonable if one considers that the sources and nature of suspended material may vary among locations, especially considering the variation in the nature of the waters sampled, and that turbidity (a measure of light scattering by suspended material) and total suspended solids (a measure of the mass of suspended materials) are two different properties of suspended materials. Consequently, a plot of the log-normalized ratio of turbidity (NTU) to total suspended solids (mg/liter) by monitoring location reveals that some sites have suspended materials with significantly different ratios of light scattering to mass than others (Fig. 3). We interpret this result to mean that the properties of suspended materials vary significantly among monitoring locations, with important implications for management of sedimentation and turbidity impacts on water quality.

Temporal variation is another factor that drives significant variability in our data set. The most obvious temporal pattern is seasonal variation in temperature and its effect on the solubility of dissolved oxygen. Our dissolved oxygen data set shows a very strong seasonal effect (Fig. 4). However, seasonal variation accounts for only a portion of the variability in dissolved oxygen values. Variability among monitoring locations is again a significant factor contributing to overall variability.

We expected that storm events would drive a significant portion of pollutant loadings to surface waters in the 201 Planning Area and that sampling results from event and non-event periods would differ accordingly. We examined the fecal coliform bacteria data set, owing to the short lifetime of fecal coliform bacteria in the environment, to look for such storm event effects and again found patterns quite different from those expected (Fig. 5). There is a tendency for fecal coliform values to be higher during or immediately after (<24 hours) a rain event, but the lowest fecal coliform values for over one third of the monitoring locations are found during a rain event. A few of the highest fecal coliform values found at some locations occurred during non-event periods. Some of the highest fecal coliform values, e.g., those at monitoring locations 3, 4, and 5, although recorded during storm events, were apparently caused by discharges of incompletely chlorinated sewage from a private sewage treatment system upstream of those locations (Fig. 6). Some of the locations with higher than average fecal coliform values, e.g., monitoring locations 7, 12, and 27, are downstream of residential areas served by septic tanks, which may overflow during rainy periods (Fig. 6). For example, Fig. 7 shows fecal coliform data during the period March 25, 1997 through May 25, 1998 for monitoring locations 20 (near a golf course), 27 (a small basin with many septic tanks) and 31 (downstream of a septage disposal site). Southeastern North Carolina experienced very heavy el Niño rains during January-February, 1998, which appeared to cause septic tanks to overflow upstream of location 27. Disposal upstream of monitoring location 31 of septage pumped from failing septic tanks in the 201 Planning Area has apparently caused fecal coliform contamination of the water at that location independently of storm event runoff (Figs. 6, 7). Consequently, a variety of human factors appear to confound the expected effects of storm events on pollutant loadings.

Golf courses also contribute pollutants to surface waters in the SBWSA 201 Planning Area. Several golf courses have been under construction while our water quality monitoring program has been underway. Construction activities at courses upstream of monitoring locations 32, 1, and 2 have apparently contributed to turbidity levels considerably higher than at most other monitoring locations (Figs. 8, 9). Golf courses also use considerable quantities of fertilizers. Our nutrient data show somewhat elevated total phosphorus values at monitoring locations 19, 20, 21, and 22, all near or just downstream of golf courses, although the highest average total phosphorus values are associated with sewage effluent discharged upstream of locations 4 and 3 (Figs. 6, 8, 10). Nutrient loading also contributes to elevated chlorophyll a levels we observe at monitoring locations 19, 20, 22, 25, and 35 (Figs. 8, 11). Monitoring locations 14 and 25, which are located on lakes near a golf course, are heavily loaded with aquatic macrophytes, another indicator of nutrient loading that is not quantified in our monitoring program. These loadings of pollutants from golf courses are attributable to stormwater runoff, inadequate runoff controls, and lack of riparian buffers to absorb pollutants.

Variability in water quality data from routine monitoring results from several factors, including among-site differences, seasonality, variable effects of stormwater runoff events, and differences in the nature of point sources. This variability can obscure patterns that might otherwise be expected. Thus the goal of establishing "baseline" or "average" conditions in any area requires a substantial data base and, in many cases, sophisticated data manipulation and statistical analyses.

The inherent variability in water quality data and the multiple sources of that variability make the task of explaining the data to the public or other users correspondingly more difficult. Such simple questions as, "Is the water quality at this site good or bad?", or "What is causing the water problems here?", become much more difficult to answer when one considers a data base large enough to show the true variability associated with any of the sites or measurements. Nevertheless, it is possible to explain what typical values for a given parameter at a given location should be when a data set is large enough, and to demonstrate what an unusual value might indicate. We have had some success in pointing out anomalies in our data that subsequently were confirmed to have been caused by unusual events in the field, such as waste spills.

Some of the variability in any given parameter can sometimes be better understood if other parameters are measured simultaneously, as impacts on water quality often affect more than one parameter. However, even our sampling protocol, which measures eleven parameters simultaneously, leaves some gaps. Consequently, we have tried to identify or devise other methods of analyzing water quality, such as analysis of detergents as a signal for human-derived wastewater and analysis of sediment-bound coliform bacteria as an indicator of longer term fecal loading patterns. One measure for which we would like to find a standard technique is macrophyte biomass, as very large portions of primary producer biomass in nutrient-loaded surface waters appear to occur in this form. These thoughts lead to the conclusion that the monitoring and research communities must continue to discuss information needs.

We thank the South Brunswick Water and Sewer Authority for its generous financial support of this work. We thank Eric Cullum, Bryant Sykes, Chris Collura, Lynn Bullard, and Zhehong Ying for their contributions to the project. Finally, we thank the duck lovers of Carolina Shores for keeping us on our toes.

Figure 7. Plot of mean fecal coliform bacteria concentrations (CFU/100 mls) between March 25, 1997 and May 25, 1998 for locations 20 (near golf course), 27 (small basin with septic tanks), and 31 (downstream of septage disposal site). The period between day 300 and day 400 had major el Niño rains.

Figure 8. Map of the SBWSA Area showing monitoring locations near golf courses.

Figure 9. Plot of turbidity (NTU) for all 36 monitoring locations during the period October, 1996 to May, 1998.

Figure 10. Plot of total phosphorus (m g/liter) for all 36 monitoring locations during
the period October, 1996 to March, 1998.

Figure 11. Plot of chlorophyll a (m g/liter) for all 36 monitoring locations during
the period October, 1996 to May, 1998.