Sound Data Analysis Procedures

ERDC TN-DOER-E14

August 2001

to a RESON EC-6070 hydrophone audio amplifier via a 50-m deck cable. The EC-6070 was used

to amplify the source levels an additional 18 dB (total source amplification of 28 dB) before the

audio data were recorded on a TASCAM DA-P1 Digital Audio Tape (DAT) recorder. All

hydrophone source audio data were recorded on the left audio track of the DAT recorder, while

simultaneous field notes were narrated and recorded on the right audio track. The hydrophone audio

data were input into a Sound Technologies analog-to-digital converter where the audio data were

digitized and stored on a laptop computer. Sound Technologies Spectra Lab v4.32 audio analysis

software was used to display a real-time audio spectrum.

The entire system was powered from two 12-V d-c deep-cycle marine batteries connected to a

StatPower pure sine-wave power inverter, which provided a 120-V a-c power source to an

uninterruptible power supply (UPS). The UPS in turn directly powered the DAT recorder, laptop

computer, and a ProTek model 3033 variable-voltage d-c power supply, which was used to provide

24-V d-c power to the EC-6070 hydrophone audio amplifier. By using the two 12-V d-c batteries

as the only power source, the entire system could be operated with the listening vessel completely

shut down to a "quiet" mode. This eliminated any noise effect that would be introduced by the

engine or generator operating on the vessel used as the listening platform.

A HydroLab water quality surveyor, recording depth, temperature, and salinity data, was suspended

1 m above the hydrophone. The hydrophone cable, HydroLab cable, and suspension line were

encased in cable fairing to prevent vibrational strumming. Depth, temperature, and salinity data

were used for speed of sound calculations (Table 1). The unit was turned off during actual audio

data recording to eliminate noise being generated by the water quality meter.

All data collection sessions were conducted aboard the Manson Dredging Company's steel hull

launch Margaret M. Because of extremely high tidal amplitudes and flow velocities in Cook Inlet,

it was determined that monitoring the acoustic levels from a fixed or anchored position was not

possible. Instead, a "drift" transect approach was used in which the Margaret M was maneuvered

to a predetermined distance away from the sound source and was then completely shut down and

allowed to drift with the current during the recording session. This approach minimized the flow

conditions present at the hydrophone, thereby reducing drag imposed on the hydrophone. During

the 1999 field study, two separate dredging operations were in progress: the bucket dredge Viking

working on the navigation channel deepening project and the smaller bucket dredge Crystal Gayle

performing maintenance dredging at the port docks.

Sound Data Analysis Procedures: Nineteen recording sessions were conducted in September

1999 (Table 1). Multiple recording sessions of ambient sounds were conducted in August 2000.

Locations of individual drift transects, determined by differential Global Positioning System

coordinates of start and end points, are depicted in Figure 1. Each recording session was digitized

from DAT tape using a Sound Technologies analog-to-digital converter and SpectraLab software

into MS Windows-compatible 16-bit stereo WAV files with a sampling rate of 44.1 kHz. Each of

the 24 WAV files was reviewed and the contents of each file summarized (Table 1). Sessions 1 and

2 were used only to determine appropriate gain settings for the hydrophone audio channel (left) on

the DAT recorder and not for audio analysis purposes. Sessions 3-9 and 11-15 taken in September

1999 monitored the Viking dredging operations and recorded dredge cycles (defined in the following

paragraph) at a number of distance ranges from the dredge. Session 10 recorded a complete dredged