The Santa Barbara Channel Experiment (SBCX)

Comparison of Radiated Noise from Commercial Container Ships and Source Signals used in SBCX

May 8, 2000

Summary: Low frequency sound radiated by commercial container ships during normal operations is hundreds of times louder than the 170 dB (1 Acoustic Watt) acoustic limit guideline used during the Santa Barbara Channel Experiment (April 1998). A comparison of an acoustic source which obeyed the limit and sound radiated from commercial ships follows.

Click on any of the illustrations for a larger version.

Our story begins in the Santa Barbara Channel, off the coast of Southern California. In April of 1998, DARPA sponsored an experiment to test new passive acoustic localization technologies. This was called the Santa Barbara Channel Experiment. A 150 element passive array of hydrophones (underwater microphones) was deployed in water 200 meters deep in the Channel. The research vessel Acoustic Explorer, towing a calibrated J-15 acoustic source, steamed around the channel area. Our job was to passively locate the ship, using only the sound radiated from the J-15 source.

The J-15 towed acoustic source was designed to radiate sound underwater. Prior to deployment, the source was calibrated at TRANSDEC, an outdoor acoustic testing facility located at SPAWAR in San Diego. Environmental regulations limited us to a total power output of 170 dB re 1uPa/1 Hz (approximately 1 acoustic Watt).

Two types of acoustic signals were played. The first was an "M-sequence," centered at 300 Hz. This was designed to test the acoustic propagation characteristics of the channel. This type of sequence could also be used for underwater acoustic communications. A 15 second sample of the actual M-sequence which was played is available here.

The second type of signal was a 12-tone "comb" signal. This was made up of 12 narrowband tones, and was designed to simulate the type of sound which would be radiated by a noisy submarine. Again, the total power output was limited to 170 dB; a short 15 second segment of the actual sequence is available here.

During the experiment, we had difficulty detecting the radiated signal from the Acoustic Explorer, due to interference from other ships. The experiment took place in the middle of an active shipping lane, which carried cargo container ships from the Pacific Ocean to and from the Ports of Los Angeles and Long Beach. A map of the shipping channel and the location of the array is shown to the left. During the experiment run denoted A/X 9 (April 14, 1998), sound from the engines and machinery of three ships prevented us from detecting the Acoustic Explorer.

The interference can be shown in the spectrogram at left. The vertical axis is time (in minutes), and the horizontal axis is Frequency, from 0 to 400 Hertz. The color scale corresponds to the intensity of the received acoustic signal, in units of decibels (dB), with a reference intensity of 0 dB equaling 1 micropascal of pressure at 1 meter from the source, normalized to 1 Hertz bandwidth. Note that sound strength decays as a function of range from the acoustic source; this is why the calibration plot above, measured 1 meter from the source, is higher (160 dB) than the spectrum at the left, which was measured at a range between 16 and 5 km from the receiver array.
A larger version of the spectrogram is available here.

From 125 to 155 minutes, the M-sequence was played. Here, the Acoustic Explorer started at a range of 16 km from the receiver array, and ended 11 km from the array. At the 11 km range, the peak of the M-sequence carrier was 90 dB, but the peak of the power from the interfering ship was 125 dB. The sound from the container ship was over 300 times as loud as the sound from J-15 towed source, measured at the receiver array. The container ship radiated sound on multiple frequencies; these were from the engines present on the ship, as well as other machinery (pumps, electrical generators, etc.) Sound from this container ship exceeded 110 dB for over 20 minutes, as it passed by the array. You can listen to a 15 second clip of the sounds (from 0 to 1 kHz) made by the three container ships below. During this time, the comb tones are played, but they cannot be heard over the sound of the ships.

Ship 1

Ship 2

Ship 3

Note: The relative amplitudes of the audio clips have been normalized to take advantage of the full dynamic range of the .wav file format. Although the comb and M-sequence .wav files sound louder than the ships, the ships recorded below were actually over 300 times louder when observed by the receiver array.

From 155 to 212 minutes, the twelve tone comb sequence was played from the J-15 transducer. The twelve tones did not become visible on the spectrogram until the very end; even there, they were difficult to distinguish from tones from other ships' machinery. The power spectral density plot shows the frequency distribution of the acoustic energy from minute 206 to 212, when the Acoustic Explorer was 5.5 km from the receiver array; the dotted vertical lines indicate where the tones were supposed to appear. One can see the presence of engine and machinery noise from other ships completely masks the tones, even when the source is close, only 5500 meters from the array.

Conclusion: Those who are concerned that DARPA and US Navy sponsored ocean acoustics experiments are adversely affecting marine life should realize that the amount of energy radiated into the water during these experiments is hundreds of times less than that generated by commercial container ships.

Last updated: 000526
Comments/Questions: pmd@mit.edu

Ships in the SBCX logo (left to right, from the top) Lockheed P-3 ("Redhook"), Acoustic Explorer, FLIP, Danny C, USS Dolphin.

	Two types of acoustic signals were played. The first was an "M-sequence," centered at 300 Hz. This was designed to test the acoustic propagation characteristics of the channel. This type of sequence could also be used for underwater acoustic communications. A 15 second sample of the actual M-sequence which was played is available here.
	The second type of signal was a 12-tone "comb" signal. This was made up of 12 narrowband tones, and was designed to simulate the type of sound which would be radiated by a noisy submarine. Again, the total power output was limited to 170 dB; a short 15 second segment of the actual sequence is available here.

	During the experiment, we had difficulty detecting the radiated signal from the Acoustic Explorer, due to interference from other ships. The experiment took place in the middle of an active shipping lane, which carried cargo container ships from the Pacific Ocean to and from the Ports of Los Angeles and Long Beach. A map of the shipping channel and the location of the array is shown to the left. During the experiment run denoted A/X 9 (April 14, 1998), sound from the engines and machinery of three ships prevented us from detecting the Acoustic Explorer.
	The interference can be shown in the spectrogram at left. The vertical axis is time (in minutes), and the horizontal axis is Frequency, from 0 to 400 Hertz. The color scale corresponds to the intensity of the received acoustic signal, in units of decibels (dB), with a reference intensity of 0 dB equaling 1 micropascal of pressure at 1 meter from the source, normalized to 1 Hertz bandwidth. Note that sound strength decays as a function of range from the acoustic source; this is why the calibration plot above, measured 1 meter from the source, is higher (160 dB) than the spectrum at the left, which was measured at a range between 16 and 5 km from the receiver array. A larger version of the spectrogram is available here.
	From 125 to 155 minutes, the M-sequence was played. Here, the Acoustic Explorer started at a range of 16 km from the receiver array, and ended 11 km from the array. At the 11 km range, the peak of the M-sequence carrier was 90 dB, but the peak of the power from the interfering ship was 125 dB. The sound from the container ship was over 300 times as loud as the sound from J-15 towed source, measured at the receiver array. The container ship radiated sound on multiple frequencies; these were from the engines present on the ship, as well as other machinery (pumps, electrical generators, etc.) Sound from this container ship exceeded 110 dB for over 20 minutes, as it passed by the array. You can listen to a 15 second clip of the sounds (from 0 to 1 kHz) made by the three container ships below. During this time, the comb tones are played, but they cannot be heard over the sound of the ships. Ship 1 Ship 2 Ship 3 Note: The relative amplitudes of the audio clips have been normalized to take advantage of the full dynamic range of the .wav file format. Although the comb and M-sequence .wav files sound louder than the ships, the ships recorded below were actually over 300 times louder when observed by the receiver array.
	From 155 to 212 minutes, the twelve tone comb sequence was played from the J-15 transducer. The twelve tones did not become visible on the spectrogram until the very end; even there, they were difficult to distinguish from tones from other ships' machinery. The power spectral density plot shows the frequency distribution of the acoustic energy from minute 206 to 212, when the Acoustic Explorer was 5.5 km from the receiver array; the dotted vertical lines indicate where the tones were supposed to appear. One can see the presence of engine and machinery noise from other ships completely masks the tones, even when the source is close, only 5500 meters from the array.