Abstract

Laurel Hill Creek is a watershed of 125 square miles located mostly in Somerset County, Pennsylvania, with small areas extending into Fayette and Westmoreland Counties. The upper part of the watershed is on the Pennsylvania Department of Environmental Protection 303(d) list of impaired streams because of siltation, nutrients, and low dissolved oxygen concentrations. The objectives of this study were to (1) estimate the annual sediment load, (2) estimate the annual nitrogen load, and (3) identify the major sources of fine-grained sediment using the sediment-fingerprinting approach. This study by the U.S. Geological Survey (USGS) was done in cooperation with the Somerset County Conservation District. Discharge, suspended-sediment, and nutrient data were collected at two streamflow-gaging stations—Laurel Hill Creek near Bakersville, Pa., (station 03079600) and Laurel Hill Creek at Ursina, Pa., (station 03080000)—and one ungaged stream site, Laurel Hill Creek below Laurel Hill Creek Lake at Trent (station 03079655).

Concentrations of nutrients generally were low. Concentrations of ammonia were less than 0.2 milligrams per liter (mg/L), and concentrations of phosphorus were less than 0.3 mg/L. Most concentrations of phosphorus were less than the detection limit of 0.02 mg/L. Most water samples had concentrations of nitrate plus nitrite less than 1.0 mg/L. At the Bakersville station, concentrations of total nitrogen ranged from 0.63 to 1.3 mg/L in base-flow samples and from 0.57 to 1.5 mg/L in storm composite samples. Median concentrations were 0.88 mg/L in base-flow samples and 1.2 mg/L in storm composite samples. At the Ursina station, concentrations of total nitrogen ranged from 0.25 to 0.92 mg/L in base-flow samples; the median concentration was 0.57 mg/L. The estimated total nitrogen load at the Bakersville station was 262 pounds (lb) for 11 months of the 2010 water year (November 2009 to September 2010) and 266 lb for the 2011 water year. Most of the total nitrogen loading was from stormflows. The stormflow load accounted for 76.6 percent of the total load for the 2010 water year and 80.6 percent of the total load for the 2011 water year. The estimated monthly total nitrogen loads were higher during the winter and spring (December through May) than during the summer (June through August).

For the Bakersville station, the estimated suspended-sediment load (SSL) was 17,700 tons for 11 months of the 2010 water year (November 2009 to September 2010). The storm beginning January 24, 2010, provided 34.4 percent of the annual SSL, and the storm beginning March 10, 2010, provided 31.9 percent of the annual SSL. Together, these two winter storms provided 66 percent of the annual SSL for the 2010 water year. For the 2011 water year, the estimated annual SSL was 13,500 tons. For the 2011 water year, the SSLs were more evenly divided among storms than for the 2010 water year. Seven of 37 storms with the highest SSLs provided a total of 65.7 percent of the annual SSL for the 2011 water year; each storm provided from 4.6 to 12.3 percent of the annual SSL. The highest cumulative SSL for the 2010 and 2011 water years generally occurred during the late winter. Stormflows with the highest peak discharges generally carried the highest SSL.

The sediment-fingerprinting approach was used to quantify sources of fine-grained suspended sediment in the watershed draining to the Laurel Hill Creek near Bakersville streamflow-gaging station. Sediment source samples were collected from five source types: 20 from cropland, 9 from pasture, 18 from forested areas, 20 from unpaved roads, and 23 from streambanks. At the Bakersville station, 10 suspended-sediment samples were collected during 6 storms for sediment-source analysis. Thirty-five tracers from elemental analysis and 4 tracers from stable isotope analysis were used to fingerprint the source of sediment for the 10 storm samples. Statistical analysis determined that cropland and pasture could not be discriminated by the set of tracers and were combined into one source group—agriculture. Stepwise discriminant function analysis determined that 11 tracers best described the 4 sources. An "unmixing" model applied to the 11 tracers showed that agricultural land (cropland and pasture) was the major source of sediment, contributing an average of 53 percent of the sediment for the 10 storm samples. Streambanks, unpaved roads, and forest contributions for the 10 storm samples averaged 30, 17, and 0 percent, respectively. Agriculture was the major contributor of sediment during the highest sampled stormflows. The highest stormflows also produced the highest total nitrogen and suspended-sediment loads.

First posted February 1, 2013