A method to adapt watershed-scale sediment fingerprinting techniques to small-plot runoff experiments

Abstract

Suspended sediment in rivers, lakes, and streams has been identified as a problem for several reasons. Watershed-scale studies are often conducted to identify primary sediment sources utilizing scale-appropriate soil fingerprinting methods. However, fingerprinting has not been attempted in small-plot runoff experiments to better understand sediment dynamics related to management practices. The objective of this study was to apply sediment fingerprinting techniques to small-plot runoff experiments to determine the relative contribution of the plot and applied sediment to sediment exported from small plots. Sediment-free water was applied to the head and sides of constructed vegetated channels in March and September of 2008. A 10 L (2 gal) sample of the runoff was collected, the water was evaporated from it, and the remaining sediment was ground and analyzed for total carbon (C). Water mixed with soil material obtained from a different part of the landscape and with a higher soil C content than that of the plot soil was applied to the head and sides of the vegetated channels, the mixture was sampled as it was applied, and the resulting runoff from the plot was sampled. These runoff samples were processed and analyzed for total C similarly to those collected when sediment-free water was introduced to the upper end of the plot. Based on differences in total C of the plot soil and the introduced soil material, a linear relationship was developed, allowing the sediment exiting the plot to be partitioned between that soil material introduced with the inflow and that soil coming from the plot bed. The sediment C content was entered into the linear equation to determine the percent plot-derived sediment in the runoff. When soil material mixed with water was introduced to the plots, on average, 20.5% of the sediment in runoff was derived from within the plot. The sediment trapping efficiency of the vegetated channels was very high (over 90%), and accounting for percent plot-derived sediment had little effect on sediment-trapping efficiency.