ASSESSMENT OF CHANGING LAND-USE PRACTICES ON BASIN SEDIMENT YIELDS AND PROVENANCE IN WESTERN NORTH CAROLINA USING MULTIVARIATE FINGER PRINTING TECHNIQUES

The negative impacts of sediment on streams of the southern Appalachians have become a significant issue in the past decade, however, there have been few attempts to quantify the impacts of land-use alterations on upland erosion. This investigation used geochemical fingerprinting techniques and sediment mixing models within a small lake basin, Fairfield Lake (0.3 sq. km), to assess the applicability of such methods to this region. Specifically, the study attempted to determine the relative contributions of sediment through time to Fairfield Lake from 2 geologic units and 4 delineated land-cover types (forests, roads, lawns, and stream-side alluvium.).

The multivariate fingerprinting and mixing model techniques used required the collection of sediment samples from upland areas and the bed of Fairfield Lake. Upland soil samples (n=108) were collected over the Whiteside Granite-Gneiss and the Tallulah Falls Formation in both forested and developed areas. In the lake, 19 cores were collected from 16 locations; 3 cores were studied in detail. Lake and upland samples were analyzed for a suite of 16 elements. Lead-210 data on the lake sediments, combined with mapped information from air photos obtained in 1963, 1975, 1988, and 2000, indicate that sedimentation rates have increased several fold during the past two decades in response to local development.

The results of linear discriminant analysis of upland soil geochemical data indicate that soils from different bedrock and land use areas can be statistically distinguished. Using the statistically selected parameters, the sediment mixing models show a change in rock and land cover source that corresponds to the increase in sedimentation rates. Also, the model indicates that anthropogenic source areas (i.e. roads and lawns) increase in sync with periods of development. The mixing model, however, does provide some unreasonable results, which may be due to elemental loss between the uplands and the lake basin, and the presence of geologic units that are not lithologically unique. This method has promise but requires modification before use as a management tool; suggested modifications are to (1) use elemental ratios and/or less soluble elements as geochemical fingerprints, and (2) map/sample geologic units by lithotype rather than formations.