Statistical properties of backscattering signals from Doppler current measurement systems

Backscattering signals from acoustic Doppler current measurement systems are assumed to obey Gaussian properties in the time domain and the energy from the Doppler to be concentrated in a narrow frequency area. This assumption makes the estimation of the Doppler shift, proportional to the ocean current, a classical signal processing problem. It is of great interest to verify that the data, does in fact obey these Gaussian assumptions, since otherwise estimation errors can result, when an estimator based on these assumptions is used. This is common practice in commercial systems. It has been noted through the use of ADCM systems that they are not always as accurate as they should have been, T. Lago et al. (1993). This could be because the Doppler signal was more complex than has been assumed. Different statistical measures have therefore been used, such as 3D spectrograms, normal probability tests, chi-square tests, skewness analysis and kurtosis analysis, to verify the signal content. In this paper, real data from four locations has been analyzed, compared and discussed, using a variety of signal analysis and statistical tools.