Discriminating silt-and-clay from suspended-sand in rivers using side-looking acoustic profilers

Abstract

The ability to accurately monitor suspended-sediment flux in rivers is needed to support many types of studies, because the sediment that typically travels in suspension affects geomorphology and aquatic habitat in a variety of ways (e.g. bank and floodplain deposition, bar morphology, light penetration and primary productivity, tidal wetland deposition in the context of sea-level rise, sediment-associated contaminants, reservoir sedimentation and potential erosion during dam removal, among others). In addition, human-induced changes to the landscape have resulted in substantially altered suspended-sediment loads (Syvitski et al., 2005). Thus, accurate monitoring of suspended-sediment flux is necessary for informed resource management of rivers. Because of this need, a variety of techniques have been developed and applied for suspendedsediment monitoring. The traditional approach in the United States, which was developed and has been used extensively by the U.S. Geological Survey (USGS), is to collect an isokinetic, velocity-weighted sample from a river cross-section, analyze the sample in the laboratory, and use water-discharge records to compute a record of suspended-sediment flux (Guy, 1969, Guy, 1970, Edwards and Glysson, 1999, Porterfield, 1972). The labor and expense associated with this traditional approach is substantial such that the number of USGS gages reporting daily records of suspended-sediment flux decreased from 364 in 1981 to 120 in 2003 (Osterkamp et al., 2004). Also, the traditional sampling approach is limited with respect to the temporal resolution that can be achieved, thus requiring the use of approximate relations between suspended-sediment concentration and water discharge to fill gaps between samples. To address these limitations, several indirect or "surrogate" measures have been investigated (see e.g. Gray and Gartner, 2009) most notably optical backscatter (i.e. turbidity), laser-diffraction, and acoustic backscatter. These indirect techniques rely on measurements of ancillary properties that correlate with suspended-sediment concentration and particle size and thus require the collection of traditional samples for calibration. Through in situ deployments, these methods can provide the high temporal resolution that cannot be achieved through traditional sampling. Here we focus on the evaluation of acoustic profiling techniques (e.g. acoustic-Doppler sideways-looking profilers, or ADPs). One major advantage of acoustic profiling is the ability to concurrently measure water velocity (using Doppler-shift methods) and suspended-sediment concentration such that suspended-sediment flux can be directly computed using data from a single instrument. Acoustic-Doppler profilers have become popular for measuring water velocity and discharge in rivers, through both moving-boat operations and from fixed deployments such as bank-mounted sideways-looking instruments (Hirsch and Costa, 2004, Muste et al., 2007). The method presented herein is most suited to sideways-looking applications as a complement to the "index velocity" technique, whereby an index velocity from a sideways-looking instrument is related to the cross-section average velocity (determined from moving-boat discharge measurements) as a means for developing a continuous water-discharge record (Ruhl and Simpson, 2005). Topping et al. (2007) presented a method for discriminating silt-and-clay from suspended sand, using single frequency ADPs. This method takes advantage of the relations among acoustic backscatter, sediment-induced acoustic attenuation, suspended-sediment concentration (SSC), and particle size distribution (PSD). Backscatter is the amount of sound scattered back and received at the transducer while sediment-induced attenuation is the amount of sound scattered in other directions and absorbed by the sediment particles. Both of these parameters can be measured with an ADP, and their different dependencies on SSC and PSD allow for the discrimination of suspended silt-and-clay from suspended sand. Topping et al. (2007) describe application of the method at several sites along the Colorado River in Grand Canyon, and herein we present an example application of the technique for the Gunnison River, CO. However, the methods general applicability in rivers has yet to be evaluated due to a lack of concurrent acoustic and sediment data at a range of sites. To this end, the objective of the analysis presented herein is to evaluate the potential general applicability of the method, drawing from the extensive USGS database on SSC and PSD. We refer to it as "potential" general applicability because it relies on the theory underlying the previous empirical results. Use of the theoretical relations is necessary due to the lack of concurrent ADP and SSC/PSD data, but also serves the additional purpose of providing further justification of the empirical calibrations developed for the Colorado and Gunnison Rivers.