Improved flow velocity estimates from moving-boat ADCP measurements

Abstract

Acoustic Doppler current profilers (ADCPs) are the current standard for flow measurements in large-scale open water systems. Existing techniques to process vessel-mounted ADCP data assume homogeneous or linearly changing flow between the acoustic beams. This assumption is likely to fail but is nevertheless widely applied. We introduce a new methodology that abandons the standard assumption of uniform flow in the area between the beams and evaluate the drawbacks of the standard approach. The proposed method strongly reduces the extent over which homogeneity is assumed. The method is applied to two field sites: a mildly curved bend near a junction featuring a typical bend flow and a sharply curved bend that features a more complex sheared flow. In both cases, differences are found between the proposed method and the conventional method. The proposed technique yields different results for secondary flow patterns compared with the conventional method. The velocity components estimated with the conventional method can differ over 0.2 m/s in regions of strong shear. We investigate the number of repeat transects necessary to isolate the mean flow velocity vector from the raw ADCP signal, discarding the influences of noise, positioning and projection errors, and turbulence. Results show that several repeat transects are necessary. The minimum number of repeat measurements needed for robust mean velocity estimates is reduced when applying the proposed method.