Examining seasonal trends in sediment source contributions in an intensely cultivated Midwestern sub-watershed using Bayesian un-mixing

摘要

This study aims to shed some light on the dynamics of sediment source contributions in the South Amana Sub-Watershed (SASW) in Iowa, USA, at a seasonal scale. Field studies were performed periodically from April to August 2007, in which sediment source and eroded samples were collected from the uplands and the stream network. The δ~(13)C and δ~(15)N signatures of the samples were obtained via mass spectrometry and used as "fingerprints" to identify the different contributing sources. For sources that were not sampled, literature values of δ~(13)C and δ~(15)N from similar systems in the region were adopted. Un-mixing of the eroded samples using the tracer signatures was achieved with an enhanced version of an existing Bayesian, Markov Chain Monte Carlo (MCMC) model, structured to well represent the intensely cultivated SASW system. Balance in sediment source contributions shifted between upland and in-streams sources over the season. Preliminary results from the study confirm observations from other studies that hydrological factors and sediment availability significantly affect the relative contributions of upland and in-stream sources. In addition, the study shows that canopy cover vs. bare soil plays an important role and should be taken into consideration when performing un-mixing studies designed to identify sediment sources.