Evaluation of an ADCP fish-bias rejection algorithm

Downward-looking surface-moored Acoustic Doppler Current Profiler (ADCP) data is at times biased towards lower values when compared to mechanical current meters (MCM). This bias has been shown to be caused by the attraction of large pelagic fish to surface moorings. RD Instruments ADCPs have an optional algorithm which attempts to limit velocity bias due to fish. The algorithm compares echo intensity amplitude among the four beams of the instrument and rejects samples in which the highest and lowest echo intensity amplitude differ by more than a given level. The performance of the algorithm was evaluated at deployments in the eastern and western equatorial Pacific. Bias was found to be reduced in some cases, but remained in others. The cause of residual bias is shown to be the presence of fish in all four acoustic beams-a condition not considered by the algorithm. A post-processing method of correcting fish bias in ADCP velocity data is presented. The method involved computing vertical modes of the bias from ADCP-MCM differences using empirical orthogonal functions (EOFs). The vertical structure of the EOFs was interpolated to the higher vertical resolution of the ADCP and then applied as a correction to the ADCP data. The quality of ADCP data was significantly improved by this correction.<>