On robust estimation of water current using a frequency domain approach: the Symmiktos method

The estimation of water current by using acoustic doppler current meters is a well-known method. Determination of the water current is based on the doppler shift caused by the movement of small particles in the water, related to the current as an average. These movements cause a doppler shift with an assumed Gaussian-shaped distribution in time. This assumption is valid for clean water and water where one particle size dominates the scattering and thereby the doppler shift. If many bubble sizes are involved in the scattering, the Gaussian distribution assumption is inaccurate. The result is a distribution where Gaussian-based estimators may generate errors. Most ADCM instruments are calibrated or verified in a basin with clean water. In consequence, they seem to be accurate, but may in fact generate errors depending on the actual water situation. In order to verify this performance, a mathematical model of the doppler shift is presented, where effects from different bubble sizes can be taken into account. This mixed statistical model is then used for synthesizing data with a different level of complexity, but with known parameters. A robust estimation method makes it possible to calibrate an acoustic doppler current meter in a basin, while relying on its data in other water situations