Joint Federal Interagency Sedimentation Conferenet and Federal Interagency Hydrologic Modeling Conference 2015

Full Program »

Large River Bed Sediment Characterization with Low-Cost Sidescan Sonar: Case Studies from Two Settings in the Colorado (Arizona) and Penobscot Rivers (Maine)

Extend Abstract
View File

Mapping subaqueous riverbed sediment grain size is difficult where and when the water flow is too swift or deep to wade yet impractical to access with large boats and instruments. Fluvial characteristics can further constrain sampling options, particularly where flow depth, water column turbidity or channel bottom structure prohibit use of aerial or bottom deployed imaging platforms.

Sidescan sonar returns that image swaths of the bed from a vessel have the potential to meet the technical shortfall confronting bed sediment change detection in large rivers. Inexpensive sonar devices designed to be mounted to small vessels and that are easy to use are commercially available. They are lightweight and have low power demands, providing opportunities for use in a large range of rivers by one or two personnel. The modern sidescan transducers are low profile and require minimal draft, making them suitable for imaging in very shallow water. Swath mapping using these devices has the potential to rapidly map bed sediments with minimal logistics and cost. Coupled with a GPS or other type of vessel tracking, they can produce geo-referenced images of the acoustic returns and relate spatial variations in the signal ('bed texture') to the grain size of the bed surface sediments.

The typical spatial resolution (pixel size) of a sonar signal return varies from decimeters to meters depending on range and acoustic parameters. The acoustic texture relates to morphological form roughness rather than the grain-scale roughness. The strength of the returned echo is a function of the bed sediment composition. A harder surface with greater acoustic impedance, such as rocks and cobbles, will return more acoustic energy than a softer bed such as sand. The predictable relation between the sonar signal, acoustic texture and substrate properties provides a basis to distinguish dominant grain sizes of a sedimentary environment. Here we discuss considerations in the use of sidescan sonar for riverbed sediment classification using examples from two large rivers, the Colorado River in Arizona and the Upper Penobscot River in northern Maine. The case studies represent two fluvial systems that differ in physiography, sediment transport, and dynamics. First, we discuss data collection 'best practices' based on experience in varied environments. Second, we relate uncertainties in instrument positioning and boat attitude (heading and pitch) to sidescan texture measurements. Third, we present methods to relate raw echoes to backscatter amplitudes (in dB Watts) and acoustic impedances by correcting for transmission, spreading and absorption losses, the sonar footprint, and instrumental factors such as time-varying transducer power. Fourth, we discuss the merits (and some pitfalls) of likely approaches to automated sediment classification from sidescan imagery, such as textural classification based on machine learning and spectral signal decomposition. Finally, we present a promising spectral technique for automated sediment classification from sidescan echograms.


Daniel Buscombe    
Flagstaff, AZ
U.S. Geological Survey

Paul Grams    
Flagstaff, AZ
U.S. Geological Survey

Theodore Melis    
Flagstaff, AZ
U.S. Geological Survey

Sean Smith    
Orono, Maine
University of Maine


Powered by OpenConf®
Copyright ©2002-2013 Zakon Group LLC