FY 1993

Fish-induced bias in acoustic Doppler current profiler data

Freitag, H.P., M.J. McPhaden, and P.E. Pullen

In Proceedings, Oceans '92, Mastering the Oceans Through Technology, Newport, RI, 26–29 October 1992, 712–717 (1992)

ADCP measurements assume acoustic reflection from plankton and other small objects flowing passively with the current. We find that this assumption is in general valid, based on comparisons of the ADCP time series with data from 5-7 companion Vector Averaging Current Meters (VACMs) and Vector Measuring Current Meters (VMCMs) on moorings in the equatorial Pacific. However, a few months after deployment, ADCP speeds at some depths and locations can be biased low relative to those of the VACM/VMCMs. The characteristics of this bias are that it is most pronounced during daylight and in the upper 100 m. These characteristics lead us to suspect that the acoustic signals are being biased by returns from pelagic fish which are known to aggregate around floating objects, and which are frequently observed around the moorings on recovery/deployment cruises. To investigate the possible impact of free swimming fish, and to determine sampling methods that might alleviate the observed ADCP speed bias, single-ping data were recorded for a 6-hour period immediately prior to recovery of a mooring at 0°, 110°W in October 1991. This was a time when the ADCP was reporting low speeds relative to a real-time VMCM near the surface. Post-recovery inspection of these data showed large beam-to-beam differences in acoustic echo intensities due to sporadic increases of up to 50 db in individual beams. Ensemble averages computed from the single-ping data showed velocity biases similar to those computed by the ADCP in the normal operational mode. Ensemble averages were then recomputed after omission of velocity values associated with high beam-to- beam echo intensity differences. This method of computation significantly reduced the bias, although some ensembles were lost due to an insufficient number of samples after editing. A modification to the ADCP firmware to perform beam intensity checks before computing ensemble averages has been developed in collaboration with RD Instruments, the manufacturer of our ADCPs. This firmware modification is currently being tested.