Verifying Sediment Fingerprinting Results with Known Mixtures
Abstract
Sediment fingerprinting is a widely used approach to determine the specific sources of fluvial sediment within a watershed. It relies on the principle that potential sediment sources can be identified using a set of chemical tracers (or fingerprints), and comparison of these source fingerprints with fluvial (target) sediment allows for source apportionment of the fluvial sediment. There are numerous source classifications, fingerprints, and statistical approaches used in the literature to apportion sources of sediment. However, few of these studies have sought to test the method by creating controls on the ratio of sources in the target sediment. Without a controlled environment for inputs and outputs, such verification of results is ambiguous. Here, we generated artificial mixtures of source sediment from an agricultural/forested watershed in Virginia, USA (Smith Creek, 246 km2) to verify the apportionment results. Target samples were established from known mixtures of the four major sediment sources in the watershed (forest, pasture, cropland, and streambanks). The target samples were sieved to less than 63 microns and analyzed for elemental and isotopic chemistry. The target samples and source samples were run through the Sediment Source Assessment Tool (Sed_SAT) to verify if the statistical operations provided the correct apportionment. Sed_SAT uses a multivariate parametric approach to identify the minimum suite of fingerprints that discriminate the source areas and applies these fingerprints through an unmixng model to apportion sediment. The results of this sediment fingerprinting verification experiment will be presented in this session.
- Publication:
-
AGU Fall Meeting Abstracts
- Pub Date:
- December 2017
- Bibcode:
- 2017AGUFMEP51A1629G
- Keywords:
-
- 1051 Sedimentary geochemistry;
- GEOCHEMISTRY;
- 1804 Catchment;
- HYDROLOGY;
- 1825 Geomorphology: fluvial;
- HYDROLOGY;
- 1862 Sediment transport;
- HYDROLOGY