Application of acoustic tomography to reconstruct the horizontal flow velocity field in a shallow river. Issue 12 (17th December 2015)

• Record Type: Journal Article
• Authors: Razaz, Mahdi; Kawanisi, Kiyosi; Kaneko, Arata; Nistor, Ioan
• Is Part Of: Water resources research. Volume 51:Issue 12(2015:Dec.)
• Page Start: 9665
• Page End: 9678
• Publication Date: 2015-12-17
• Subjects: shallow acoustic tomography -- inverse analysis -- L‐curve -- Tikhonov regularization -- flow visualization -- ADCP; Hydrology -- Periodicals; 333.91
• DOI: 10.1002/2015WR017102
• Languages: English
• ISSNs: 0043-1397

Abstract

Abstract: A novel acoustic tomographic measurement system capable of resolving sound travel time in extremely shallow rivers is introduced and the results of an extensive field measurements campaign are presented and further discussed. Acoustic pulses were transmitted over a wide frequency band of 20–35 kHz between eight transducers for about a week in a meandering reach of theBāsen River, Hiroshima, Japan. The purpose of the field experiment was validating the concept of acoustic tomography in rivers for visualizing current fields. The particular novelty of the experiment resides in its unusual tomographic features: subbasin scale (100 m × 270 m) and shallowness (0.5–3.0 m) of the physical domain, frequency of the transmitted acoustic signals (central frequency of 30 kHz), and the use of small sampling intervals (105 s). Inverse techniques with no a priori statistical information were used to estimate the depth‐average current velocity components from differential travel times. Zeroth‐order Tikhonov regularization, in conjunction with L‐curve method deployed to stabilize the solution and to determine the weighting factor appearing in the inverse analysis. Concurrent direct environmental measurements were provided in the form of ADCP readings close to the right and left bank. Very good agreement found between along‐channel velocities larger than 0.2 m/s obtained from the two techniques. Inverted quantities were, however, underestimated, perhaps due to vicinity of the ADCPs to the banks and strong effect of river geometry on the readings. In general, comparing the visualized currents with direct nodal measurements illustrate the plausibility of the tomographically reconstructed flow structures. Key Points: Array of tomographic sensors is deployed in a 270x100m domain of a shallow river Horizontal flow field is reconstructed using 15 paths and inversion techniques Comparison with stationary ADCPs denotes the plausibility of reconstructed flow