Saturday, 15 July 2006

Sediment Source Identification in an Urban Watershed.

Olivia H. Devereux1, Brian Needelman1, Karen L. Prestegaard2, Allen Gellis3, and Jerry Ritchie4. (1) Univ of Maryland, 0104 H.J. Patterson Hall, College Park, MD 20742, (2) Univ of Maryland, CMPS-Geology, 0214A Chemistry Bldg, College Park, MD 20742, (3) USGS, Water Resources Division, 8987 Yellow Brick Road, Baltimore, MD 21237, (4) ARS-BARC-HRSL, 10300 Baltimore Ave, Beltsville, MD 20705

Suspended sediment is a water quality problem in the Chesapeake Bay. The objective of this project was to identify sediment sources in an urban watershed, the Northeast Branch of the Anacostia River (in Washington, D.C. and Maryland: drainage area = 188.5 km2), which delivers sediment directly to the Chesapeake Bay. This watershed spans two physiographic regions—the Piedmont and Coastal Plain. Bank sediment and suspended sediment deposits were characterized using the following techniques: radionuclide (Cs-137) by gamma ray spectrometry, trace-element analysis by inductively coupled plasma mass spectrometry, clay mineralogy by x-ray diffraction, and particle-size analysis with a laser particle-size analyzer. Sampling of bank and suspended sediments was designed to: a) characterize tributary inputs from both Piedmont and Coastal Plain sources, and b) differentiate tributary inputs from bank erosion along the main stem of the Northeast Branch. Thirteen sample sites were chosen that represent tributary source areas of each physiographic region and the main stem where mixing occurs. Surface samples of the banks were compared to overbank deposits from a ten-year storm and suspended sediment collected during major storm events over a one-year period. Fingerprint components were selected from the source sample data and the overbank deposit data. The fingerprint components were selected as those present in large quantities or with a coefficient of variation greater than 50. Composite fingerprints were created and statistically verified. Two methods were compared for creating the composite fingerprints—principal component analysis and multivariate discriminate function analysis. A mixing model differentiated the relative contributions of potential suspended sediment sources for each storm event. Corrections for differences between particle sizes were made. Erosion and storage of sediment patterns are discernible using preliminary Cs-137 data. Preliminary trace-element data indicates that stream banks may be differentiated from overbank deposits. This method may be applicable to other urban watersheds.

Back to 1.0WA Soil Geochemical Patterns at Regional, National, and International Scales - Poster
Back to WCSS

Back to The 18th World Congress of Soil Science (July 9-15, 2006)