Surface-water quality in agricultural watersheds of the North Carolina Coastal Plain associated with concentrated animal feeding operations

Abstract

The effects of concentrated animal feeding operations (CAFOs) on water quality were investigated at 54 agricultural stream sites throughout the North Carolina Coastal Plain during 2012 and 2013. Three general watershed land-use types were examined during the study, including 18 background watersheds with no active CAFOs (BK sites), 18 watersheds with one or more active swine CAFOs but no poultry CAFOs (SW sites), and 18 watersheds with at least one active swine CAFO and one active dry-litter poultry CAFO (SP sites). The watershed drainage areas for these 54 stream sites ranged from 1.2 to 17.5 square miles. Conventional fertilizers used for crop production are the primary source of nutrients at the BK sites. Animal-waste manures represent an additional source of nutrients at the SW and SP study sites. Land cover, soil drainage, and CAFO attributes were compiled for each watershed. Water-quality field measurements were made and samples were collected at the 54 primary sites during 6 bimonthly sampling periods from June 2012 to April 2013. An additional 23 secondary sites were sampled once during April 2013 to provide supplemental data at stream locations directly adjacent or in close proximity to swine CAFOs and (or) background agricultural areas within 9 of the primary watersheds. The watershed drainage areas for the 23 secondary sites ranged from 0.2 to 8.9 square miles. Water temperature, specific conductance, dissolved-oxygen concentration, and pH were measured directly in the streams. Water samples were analyzed for major ions, nutrients, and stable isotopes, including delta hydrogen-2 (δ 2 H) and delta oxygen-18 (δ 18 O) of water and delta nitrogen-15 (δ 15 N) and δ 18 O of dissolved nitrate plus nitrite. Most of the water-quality properties and constituents varied significantly among the six sampling periods, changing both seasonally and in response to hydrologic conditions. The differences noted among the sampling periods indicate that the interactions between seasonal climatic differences, streamflow conditions, and instream biotic and abiotic processes are complex and their integrated effects can have varying degrees of influence on individual nutrients. Water-quality differences were noted for the SW and SP land-use groups relative to the BK group. Median values of specific conductance, several major ions (magnesium, sodium, potassium, and chloride), and nitrogen fractions (ammonia plus organic nitrogen, ammonia, nitrate plus nitrite, total nitrogen, and δ 15 N of nitrate plus nitrite) were higher for the SW and SP groups compared to the BK group. No significant differences in water temperature, dissolved oxygen, calcium, total organic nitrogen, orthophosphate, total phosphorus, or δ 18 O of nitrate plus nitrite were noted among the land-use groups. When compared on the basis of land-use type, there was an overall measurable effect of CAFO waste manures on stream water quality for the SW and SP watershed groups. Some individual sites within the SW and SP groups showed no measurable CAFO effects on water quality despite having CAFOs present upstream. An evaluation of sodium plus potassium concentrations coupled with δ 15 N values of nitrate plus nitrite proved valuable for distinguishing which SW and SP sites had a water-quality signature indicative of CAFO waste manures. Sites with CAFO manure effects were characterized by higher sodium plus potassium concentrations (commonly between 11 and 33 milligrams per liter) and δ 15 N values of nitrate plus nitrite (commonly between 11 and 26 parts per thousand) relative to sites reflecting background agricultural conditions, which commonly had sodium plus potassium concentrations between 6 and 14 milligrams per liter and δ 15 N values of nitrate plus nitrite between 6 and 15 parts per thousand. On the basis of the results of this study, land applications of waste manure at swine CAFOs influenced ion and nutrient chemistry in many of the North Carolina Coastal Plain streams that were studied. A classification tree model was developed to examine relations of watershed environmental attributes among the study sites with and without CAFO manure effects. Model results indicated that variations in swine barn density, percentage of wetlands, and total acres available for applying swine-waste manures had an important influence on those watersheds where CAFO effects on water quality were either evident or mitigated. Measurable effects of CAFO waste manures on stream water quality were most evident in those SW and SP watersheds having lower percentages of wetlands combined with higher swine barn densities and (or) higher total acres available for applying waste manure at the swine CAFOs. Stream water quality was similar to background agricultural conditions in SW and SP watersheds with lower swine barn densities coupled with higher percentages of wetlands or lower acres available for swine manure applications. The model provides a useful tool for exploring and identifying similar, unmonitored watersheds in the North Carolina Coastal Plain with potential CAFO manure influences on water quality that might warrant further examination.