Fluorescence-based source tracking of organic sediment in restored and unrestored urban streams

Abstract

Fine sediment is a leading cause of impairment of waterways, but its dynamics are generally poorly understood. Most fingerprinting studies focus on the mineral component of fine sediment, yet the organic fraction has distinct effects on ecosystem processes. Here we employed a novel combination of fluorescence spectroscopy and end-member mixing analysis (EMMA) to distinguish between sources of particulate organic material (POM) in streams with and without restoration structures, situated within urban watersheds of the Chesapeake Bay. Suspended sediment samples and potential source end-member samples were leached and analyzed for their fluorescence excitation-emission (EEM) spectra. EEMs were resolved into fluorescent components using parallel factor analysis (PARAFAC). Six conservative components comprised the inputs to EMMA. The final model distinguished seven categories of source contributions: old and recent banks, the streambed, impervious surfaces, forested soils, algae, and a general category of allochthonous contributions from the watershed. The poorly characterizable allochthonous category dominated all samples, particularly at baseflow. During storms, the stream with restoration structures transported larger concentrations of fluorescent POM than the unrestored stream, with greater contributions from banks and soil. However, the bed contribution was greater for the unrestored stream, reflecting greater instream storage of fine sediment during interstorm periods. Relative source contributions in the model were insensitive to variability in the precise spectral characteristics of source end members or the number of end members, and varied in a manner consistent with external measurements, providing confidence in the approach.