J.M. Rice, Extension Specialist

Department of Biological and Agricultural Engineering

North Carolina State University, Raleigh, NC

J. Spooner, Extension Specialist

Water Quality Group in the Department of Biological and Agricultural Engineering

North Carolina State University, Raleigh, NC

M.G. Cook, Professor Emeritus

Department of Soil Science

North Carolina State University, Raleigh, NC

K.C. Stone, Agricultural Engineer

USDA Agricultural Research Service, Coastal Plains Soil

Water and Plant Research Center, Florence, SC

S.W. Coffey, Environmental Specialist

Division of Soil and Water Conservation

Department of Environment and Natural Resources, Raleigh, NC

F.J. Humenik, Coordinator of Waste Management Programs

College of Agriculture and Life Sciences

North Carolina State University, Raleigh, NC

P.G. Hunt, Soil Scientist and Research Leader

USDA Agricultural Research Service

Coastal Plains Soil, Water, and Plant Research Center, Florence, SC

Agricultural non-point source pollution has been of concern, particularly where intensive operations exist near environmentally sensitive waters. In 1989 the United States Department of Agriculture (USDA) addressed these concerns by funding eight Water Quality Demonstration Projects with the goal of increased voluntary adoption of agricultural best management practices (BMPs). The Herrings Marsh Run Demonstration Project in Duplin County, North Carolina has been able to document water quality improvements as a result of widespread BMP implementation. By combining the resources and expertise of various federal, state and local agencies and a receptive agricultural community, BMPs have been installed throughout the watershed to address several aspects of farm management and rural land use.

Accurate land use data, that could be linked to the subwatershed water quality monitoring stations, proved to the most difficult information to collect. Current government databases typically do not contain detailed records. Farmers, and in some cases contract growers, consider some information to be proprietary. Standard government records were supplemented with farm surveys and drive-by field observations to document land use changes.

Water quality monitoring and USGS gaging stations were installed on streams in subbasins within the watershed to help characterize the overall water quality of the drainage area. In addition, by utilizing subbasins it was possible to compare and contrast changing land use patterns on the watershed. Water monitoring data from the watershed outlet have confirmed an improvement in water quality as evidenced by a 50% decrease in the nitrate-nitrogen (NO₃-N) concentration.

Even though significant progress has been made in the development and implementation of agricultural best management practices (BMPs), there has been little progress at documenting the water quality impacts of their implementation on a watershed scale. A United States Department of Agriculture (USDA) Water Quality Demonstration Project was initiated in 1990 in the Herrings Marsh Run (HMR) Watershed, located in the Northeast Cape Fear River Basin in Duplin County, North Carolina. This project was one of the original eight Demonstration Projects funded by USDA as part of the 1989 Presidential Water Quality Initiative. These projects were conceived as cooperative efforts by the Cooperative Extension Service, the Natural Resources Conservation Service, and the Farm Services Agency. The overall goal of the water quality projects was to promote wide spread, voluntary adoption of BMPs to protect and improve water quality. For the North Carolina project, the USDA-Agricultural Research Service was added to the cooperative effort to provide water quality monitoring capability so that the water quality impacts of the BMPs could be assessed.

Duplin County has the highest agricultural income of any county in North Carolina and also ranks as one of the highest poultry and swine producing counties in the United States (NC Agricultural Statistics). The county is typical of the southeastern Coastal Plain of the United States with sandy soils and relatively shallow water tables. The lateral flow of shallow ground water is the primary source of base stream flow in the region.

Grab samples and observations made before the project was initiated indicated higher than expected ammonia and nitrate levels in certain sections of the watershed. Therefore, nitrogen was targeted as the primary pollutant for reduction.

The total area of the Herrings Marsh Run Watershed is 5050 acres. Agricultural management practices on the watershed are typical of the southeastern Coastal Plain and include approximately 2700 acres of cropland, 1750 acres of woodlands and 525 acres of farmsteads, poultry facilities, and swine facilities. The remainder of the land area is in road right-of-way and ponds and lakes. The major agricultural crops on the watershed include tobacco (324 acres), corn (1026 acres), soybean (650 acres), wheat (300 acres) and vegetables (400 acres). The predominant soil series in the watershed is Autryville fine sand; secondary soil series are Norfolk loamy sand, Marvyn-Gritney soil complex and Blanton sand.

BMPs targeted for implementation included many traditional practices such as conservation tillage and other erosion control measures as well as animal waste storage structures. There was also an emphasis on implementing new and innovative practices such as turkey mortality composting and nutrient management and waste utilization, both of which were based on soil type and realistic crop yields. In addition to these new management practices, there were also landscape modifications, i.e., a riparian area restoration and an in-stream wetland created by a beaver dam.

Many approaches were employed to monitor changing land management and treatment practices. The land use information for those individuals who received technical or financial assistance was available from the standard USDA records which are linked to farm and field numbers. Unfortunately, in many cases, these records are not in electronic format, that could easily be geographically referenced to a watershed or hydrologic unit. This required manual sorting of data and digitizing of maps to enable the information to be incorporated into a geographic information system (GIS). The GIS was used to facilitate tracking the changing land use patterns on a watershed basis and relating any changes to changes in water quality. To supplement these existing databases, on-farm surveys were conducted with area producers. This required a time commitment from the producers to meet with a project technician and fill out a rather detailed survey instrument. Finally, semi-annual windshield crop surveys were conducted. This consisted of touring the watershed and noting on a map the crops in each field and the presence of new structural practices. In the absence of better information, regarding crop inputs, average rates of fertilizer and pesticide application and standard management practices were assumed.

Surface water sampling stations were established in August 1990, at three locations within the watershed (Figure 1). Station 1, Red Hill, was located at the stream outlet from the watershed. Station 2 was located along a tributary downstream from intensive swine and poultry operations. Station 3 was located along the main channel flowing through woodlands. The woodlands located above Station 3 have more substantial riparian buffers compared to those buffers on the streams above Station 2. Station 3 was chosen to represent background conditions due to the large riparian buffers and relatively small areas of agricultural production. Station 4 was installed in August 1991, to provide additional information about the eastern portion of the watershed. The U.S. Geological Survey in Raleigh, NC, installed gaging stations at the initial three sampling stations in April 1991, and the final installation was completed at Station 4 in August 1991. At the gaging stations, stage height is measured at 15-minute intervals using automated water level recorders and the flow was calculated from the stage-discharge curves developed for each stream reach.

Automated water samplers installed in 1990 at each of the sampling stations were programmed to collect daily, time-based daily composite samples. The automated samplers were reprogrammed in October 1993, to collect 2-day composite samples comprised of 24 sub-samples taken at 120-minute intervals. Beginning in November 1994, 3.5-day composite samples were collected. Each composite sample consisted of 42 sub-samples collected at 120-minute intervals. Later, in March 1997, the samplers were programmed to collect a 7-day composite sample consisting of 42 sub-samples taken at 240-minute intervals. Diluted sulfuric acid was placed in the sampler bottles prior to sample collection to reduce nutrient degradation. The acidified samples were collected each week for nutrient analysis. The sample collection has been continual from October 1990 to the present time. All water quality data was converted to weekly averages to provide a consistent data set.

The North Carolina Department of Environment and Natural Resource, Division of Water Quality, investigated the macroinvertebrate biotic index, taxa richness, habitat variables and site conditions at the watershed outlet and a nearby-unrelated watershed to produce a bioclassification for the watershed. The bioclassification rating scale is: Good, Good-Fair, Fair, Fair-Poor, and Poor. The sites were evaluated annually to assess changes.

A network of monitoring wells was established on a grid system throughout the watershed to assess the general condition of shallow ground water. These wells were sampled monthly and analyzed for nutrients and selected pesticides known to be used in the area. A residential well screening program was also initiated to increase awareness of water quality issues among watershed residents and to assess the status of drinking water quality.

Most producers in the watershed were willing to consider at least a singular BMP or a system of BMPs. The use of traditional soil conservation and erosion control practices such as conservation tillage, field borders, and grassed waterways has increased significantly. Relatively new practices were also employed to reduce nitrogen sources and transport. These practices included nutrient management, waste utilization, riparian area restoration, and in-stream wetlands.

Project personnel have developed nutrient management plans for nearly 2,500 acres (over 90% of the cropland in the watershed). The amount of cropland being managed under of nutrient management plan has increased significantly since the inception of the project (Figure 4). Producers have certified that plans covering half of the acreage (1360 acres) are being followed. The status of the remaining plans is uncertain due a lack of owner certification of adherence to their plan. It is hoped that at least some parts of the plans are being followed. Several animal waste utilization plans were also developed following the same principles of nutrient application; i.e., soil testing, matching crop nutrient needs to soil test results, crediting of nutrients from animal waste and timing of nutrient application to meet the requirements of the plant. The increase in the acres associated with waste utilization is a result of larger land application areas on existing farms and more new swine facilities which began operating during the past seven years. Theses new facilities have resulted in a doubling of the swine population in the watershed during the time span of the project (Figure 5). The increase in the swine population was most noticeable in Subwatershed 3 which previously had relatively little swine production.

While the linkage between land use and improved water quality is difficult to establish with the available data, the cumulative effects of BMP implementation and changing landscape features have resulted in a decrease in the NO₃-N concentration and mass flux leaving the watershed outlet (Figures 2 and 3).

The reduction of NO₃-N concentration at Station 2 appears to be a direct result of a beaver dam impoundment since the reduction corresponds with first signs of beaver construction activity (Figure 6). While there are seasonal fluctuations, there has been a decrease in the mean NO₃-N concentration at Station 2.

The mean nitrate-N concentration at Station 3 has remained relatively unchanged even though the subbasin monitored by this station has experienced the most significant increase in swine population. This consistent water quality may reflect the adherence to approved practices and the presence of significant riparian buffers in the subbasin. Due to the relative short amount of time since animals have been introduced at the new swine facilities, the water quality results are inconclusive at this point.

The bioclassification of the watershed improved from Fair to Good-Fair following the 1995 evaluation (Lenant and Eaton, 1995). It is uncertain if this change was a response to BMP implementation or the result of annual weather variations. The severe damage caused by Hurricanes Bertha and Fran during 1996 caused the bioclassification of the watershed to decline from the Good-Fair rating to a Poor rating following the 1996 evaluation. The watershed was not re-evaluated during 1997 due to a lack of resources in the state agency doing the evaluation.

The ground water NO₃-N concentration was below the safe drinking water standard of 10 parts per million (ppm) on 78 percent of the farms which had monitoring wells installed (Hunt et al., 1995). Only five farms had wells with NO₃-N concentrations exceeding the drinking water standard. One of these farms, a swine farm, was then selected for more intensive study of Bmp implementation and water quality impacts. The new management practices and increased land area for waste application have resulted in a decrease in the average ground water NO₃-N concentration at this site (Gale et al., 1996). When pesticides were detected in monitoring wells, the concentrations were usually lower than the maximum contaminant level (MCL) set by the U.S. Environmental Protection Agency (US-EPA). The incidence of higher concentrations appear to be associated with the mixing of chemicals and their loading in chemical application equipment.

Residential drinking water wells tended to have a higher incidence of both high (>10 ppm) NO₃-N and pesticide detections. The residential wells tended to be older wells of poor construction relative to the monitoring well. Of the wells less than 100 feet deep, 27 percent had NO₃-N concentration above 10 ppm while none of the wells greater than 100 feet exceeded that level. Pesticides were detected in 34 percent of the wells with two wells exceeding the MCL. These two wells were shallow (<30 feet deep) and had historically been used to mix pesticides. Both of the wells were replaced with new, properly constructed wells greater than one hundred feet deep.

Although it has not been possible to directly link individual BMPs with improvements in water quality, on the watershed basis, the cumulative impacts of BMPs and landscape modifications have resulted in a decrease in nitrate concentration and mass loading at the watershed.

The importance of landowner cooperation and commitment in a successful watershed project cannot be over emphasized. Producers were most willing to adopt practices requiring the least risk or which could be phased into their operations. Nutrient management based on soil type and yield potential was one such practice that producers often initiated on a small acreage the first year, then expanded the use in subsequent years as they gained confidence in the results. In some cases, such as animal waste management, producers adopted practices they hope will preclude additional regulations or will enable them to remain in compliance with changing requirements.

Data indicate that, in general, current pesticide BMPs used by local producers are successful in maintaining ground water quality. The more frequent detection of nitrates and pesticides in residential wells as opposed to monitoring wells indicates that improper well construction and use contribute to the risk of contamination.

To document water quality improvements from BMP implementation on a watershed basis, it is preferable to:

• Have a commitment up-front from producers to maintain detailed records.
• Design land treatment tracking and water quality monitoring schemes that can be matched on hydrologic and temporal bases.
• Select a small watershed with the aim of reducing the time needed to detect changes in water quality.
• Isolate natural changes in a watershed, either gradual (beaver ponds) or catastrophic (hurricanes), which can mask improvements in water quality brought about by BMP implementation.
• Employ multiple approaches to monitoring in order to provide the best chance of obtaining meaningful results.
• Collect baseline water quality data to document the magnitude of change.
• Utilize a paired watershed design in which two similar watersheds are monitored for approximately 2 years to determine their relationship to each other (calibration period) and then BMPs are implemented in one of the watersheds (treatment watershed). A change in the relationship between the treatment and the control watershed indicates the water quality impact for the BMPs.