A Model Approach To Predicting Errors In Acoustic Dopler Current Profiles

Abstract

A modeling approach is applied to predicting errors in velocity estimates made by pulse to pulse incoherent acoustic Doppler current profilers (ADCPs). A system model is described that provides a direct simulation of ADCP signal processing, including a tracking feedback loop, sometimes referred to as a frequency tracker. Although the model follows general principles of incoherent sonar signal processing (autocovariance estimation of the Doppler shift frequency) and phaselocked loops, we present model predictions for a particular system, an RD Instruments 153.6 kHz ADCP, a system in widespread use on oceanographic research vessels and current meter moorings. Model validation is achieved by comparison with ADCP measurements of the same signals, coupled into the ADCP using a signal simulator.By assuming a statistical model of the backscatter, the model can be used to predict errors in velocity estimates determined by Janus beam pairs, measuring a velocity shear in a medium of randomly distributed scatterers. In addition to error prediction, the model is used to explore the dependence between velocity errors and system parameters, with an emphasis on optimizing those parameters that are accessible to users of commercial instruments.