Occurrence and distribution of mercury in streams and reservoirs in the Triangle Area of North Carolina, July 2007–June 2009

During the time period 2001–2006, the U.S. Geological Survey reported mercury-concentration measurements that exceeded the North Carolina water-quality criterion (NCWQC) of 0.012 microgram per liter for total recoverable mercury in streams and reservoirs across the Triangle Area of North Carolina. Mercury data were sparse, however, generally consisting of only one or two water samples per year. Additional monitoring and data analysis were needed to better determine the occurrence and distribution of mercury in the Triangle Area for all seasons and waterbody types as well as associations between mercury concentrations and water-quality and land-use parameters. Water at fifteen reservoir and 14 stream sites across the Triangle Area was sampled at various times between August 2007 and June 2009, with water samples collected from both the surfaces and bottoms of the water columns in reservoirs and from the surfaces of streams. A bed sediment sample was also collected at all reservoir sites and at all but one stream site. A total of 301 water samples was collected at reservoir sites. Filtered and total recoverable mercury were detected in at least one water sample collected from each reservoir site. A total of 77 water samples was collected from stream sites with filtered mercury detected in samples from one-half of these sites, and total recoverable mercury detected in at least one water sample from all but two sites. Total recoverable and filtered mercury concentrations exceeded the NCWQC for mercury more frequently in reservoir than in stream samples. Differences in sampling frequencies among seasons and between streams and reservoirs, however, may have negatively biased overall estimates of mercury concentrations in streams relative to reservoirs. Filtered mercury concentrations in surface-water samples from reservoirs and total recoverable mercury concentrations in bottom samples from reservoirs were highest in the fall, whereas no seasonal trends in filtered or total recoverable mercury were detected from stream samples. Total mercury concentrations were calculated for the bulk sample on the basis of the percentage of the grains in the bulk sample whose diameters that were smaller than 0.0625 millimeters. Total mercury concentrations in bed sediment were generally higher for samples from reservoir sites compared to streams sites, although the highest total mercury concentration in bed sediment was from a stream site. Concentrations of total recoverable mercury in water samples from stream sites all fell within the general range for streams and lakes without on-site significant anthropogenic sources (for example, mercury mines or industrial pollution), whereas samples collected from eight reservoir sites had total mercury concentrations in a range characteristic of sites affected by mercury mines or industrial pollution. Results suggested that litterfall may be a source of mercury in streams, whereas atmospheric deposition is likely a dominant source for reservoirs; however, high concentrations of filtered and total recoverable mercury concentrations in the fall season in some reservoir-water samples may warrant further analysis of potential hydrologic factors. Mercury concentrations in all water and bed sediment samples were below levels expected to cause adverse effects to humans and aquatic biota, indicating that mercury levels at the study sites in the Triangle Area were unlikely to cause an immediate health risk to humans or aquatic organisms. The high variability among several sample replicates for total recoverable mercury, however, indicated that inferences from total recoverable mercury concentrations can be tenuous.