Fifty miles off the California coast, just over the edge of the continental shelf, an underwater mountain rises from the Pacific Ocean floor, cresting out 900 meters below the ocean surface. This underwater aerie, half again as high as Mount Diablo, surveys the open ocean to the west, the Juan de Fuca Plate to the north, and, to the south, the teeming wildlife of the Monterey Bay National Marine Sanctuary. Pioneer Seamount is an ideal vantage point for seeing everything that goes on on one side of the North American continent.
But seeing underwater is not done with light. In even the clearest of seawater, light is strongly absorbed - two-thirds of blue light is absorbed over a distance of fifty meters, and red light fares worse. In the murkier water of the Pacific, a whale can rarely see its own tail, much less its mate or a straying baby. In this situation, sound replaces light and ears become eyes.
In contrast to light, sound underwater travels almost forever. Frequencies of 50 Hz and below, favored by some whales, travel half the circumference of the earth before attenuation sets in. Higher frequencies are attenuated more quickly, but dolphin echolocation frequencies of 50,000 Hz still have an attenuation length of over a kilometer.
Another factor favors detection of distant sounds from Pioneer Seamount - the underwater sound channel. The combined effects of warmer temperatures at the surface and greater pressure at greater depths result in a minimum in the speed of sound; off the California coast this minimum is found 600 m below the surface. This variation in sound speed produces a focusing effect on sound waves propagating near this depth, keeping the sound near the channel axis. (This sound channel was first discovered by military acousticians, and was referred to as the SOFAR (Sound Frequency and Ranging) channel.)
The advantages of the sound channel for long-distance communication led the scientists of the Acoustic Thermometry of Ocean Climate (ATOC) project in 1995 to choose Pioneer Seamount as a site for transmission and reception of low-frequency signals. Sound travel times between the California and Hawaii transmitters and a number of receiving stations were to be used to measure the average temperature of the Pacific Ocean, providing essential data for evaluating the role of oceanic heat storate in global climate change. A coaxial submarine telephone cable was laid from Pillar Point, near Half Moon Bay, to the underwater site, and a high-power acoustic source was installed. This source transmitted a frequency-coded signal in a band near 75 Hz, at source levels of about 212 dB re 1 mPa at 1 m. (This corresponds to about 200 Watts of acoustic power, similar to the sound system of a band performing on stage.) The frequency band overlaps the frequency spectrum of blue whale calls (and those of other species, to a lesser extent), and the contentious issue of possible effects on these and other marine animals was never satisfactorily resolved. After five years of operation, this source was turned off, and the ATOC team prepared to remove the cable.
At this time, an initiative to preserve the cable for use in non-invasive environmental monitoring was undertaken, spearheaded by oceanographer Newell Garfield, with the support of David Evans, Director of NOAA Ocean and Atmosphere Research (OAR), and concerned environmental groups acceded to the logic of this proposal. A team of scientists led by Chris Fox of NOAA's Pacific Marine Environmental Lab (PMEL) and Jim Mercer of the University of Washington's Applied Physics Lab (APL) then installed a small vertical linear array (VLA) of four hydrophones and equipment to collect the signals from these hydrophones, 24 hours a day, archive them and make them available to the public over the internet. On September 1, 2001, the Pioneer Seamount Observatory came on line.
During the following year, the observatory suffered a variety of minor equipment problems, and one failure which required bringing the "wet electronics" to the surface for repairs. This entailed a wait of four months for ship availability and suitable weather conditions. Still the Observatory's live time averaged nearly 60% during a period of over a year, and a large body of data is now available for analysis.
The accompanying figure is a composite spectrogram of acoustic signals commonly observed at Pioneer Seamount.
The spectrograms use a windowed Fourier transform to show "frequency versus time," and most of the interesting phenomena can be located by scanning the spectrograms. Four signals of interest are marked on the spectrogram.
Ship-Propeller Sounds. The most obvious and loudest feature is the pattern of concentric parabolic lines covering most of the spectrogram. This is the signal of a ship passing over Pioneer seamount. Not only is the shipping traffic the loudest single noise recorded at PSM, it is also a major contributor to the ambient noise level, in other words, the noise generated by shipping traffic travels long distances. This complex pattern is generated by the interference between the four hydrophones of the VLA, whose signals are added coherently. The point where the lines dip down to their lowest frequencies is the point of closest approach for a passing ship, and the frequency at that point gives the distance of closest approach of the ship to a point directly over the VLA. From the rate at which the interference lines diverge, the speed of the ship can be determined. The pattern shown corresponds to a ship passing about 300 m from the array's location, and traveling roughly in a straight line, at a constant speed of 12 knots.
Not all of the ship signals are so simple. If the ship changes course, the lines bend. Reflection of sound from the bottom and from topographic features near the VLA produce distortions in the interference pattern. Analysis and computer simulation of these patterns is in progress.
Blue-Whale Calls. On the lower left-hand side of the figure appear a series of four blue-whale (Balenoptera musculus) "A-B" calls. Each pair starts with an "A" call about 20 seconds long, with substantial power at 16 Hertz (below the limit of human hearing) and at 95 Hz, the fifth harmonic of the low-frequency fundamental. The "B" call follows about 60 seconds later, and has its frequencies concentrated at 16 and 48 Hz, the first and third harmonics of the same fundamental. The sounds of blue and fin whales are generally played back at between four and ten times their true speed, moving the sound frequencies into the center of the human range of hearing. The "A" call sounds like a series of "gurgles," and the "B" call which follows sounds like a sad "moan," dropping steadily in frequency during its 15-second duration.
The "B" call, the less complex of the two, turns out to be fairly easy to recognize with automated pattern recognition. An effective method described in the literature uses a "matched filter" consisting of a perfect sine wave at about 16 Hz, dropping slightly in frequency during the "moan." This simplified version of the call is used as the kernel for cross correlation with the time series (sampled at 1000 samples/sec). This filter, with a threshold set well above the noise, detects blue-whale calls down to a level where they are difficult to discern from visual inspection of the spectrogram. This work is preliminary, but the rate of false positives is less than 20%, coming mainly from loud shipping noises feeding through into the frequency band of the "B" call.
In 270 days of live time we detected about 15,000 blue-whale "B" calls. The frequency of these calls is shown in figure 2, with the strength of each call plotted in part (a) and the calls per week in part (b). While the full year is not available, the difference between the busy fall and a vacant spring is striking.
The large body of whale calls tempts one to speculate about their meaning. Marine mammals are thought to be among the most intelligent life forms on Earth. Sound is their primary means of communication. What do these calls mean? One interesting form of variability is the presence or absence of the "A" call. There are several occurrences in this data set of a whale giving a series of "B" calls, without an accompanying "A" call. (The "A" call is clearly distinguished by the strong component at about 90 Hz, as seen in the spectrogram.) The recurrence of this variant in the PSM data may eventually lead to finding an interpretation for this behavior. The "holy grail" of marine-mammal acoustics is the identification of individuals from their calls. If blue whales could be identified from their "A-B" calling sequence, one could study their migration patterns, their family groupings, and any other behavior bringing them regularly near PSM, with all the advantages of a long continuous data set. However, blue-whale calls contrast dramatically with those of the humpback - while the humpbacks soar wildly from high to low tones in a very complex pattern, the blue whale has just two sounds - the "gurgle" and the "moan" - repeated over and over again with remarkable little variation. Still there are a number of minor variations which may provide the key to identifying individuals. The distribution of power in the call among the fundamental frequency and its overtones (frequency multiples) shows considerable variability. We are just beginning to study the distribution in central frequency and rate of change of frequency for the "B" calls, and these data may provide some way to tag individuals, age groups or sex groups. This is a promising direction for future research at Pioneer Seamount.
RAFOS timing sources. The fine, nearly horizontal line on the spectrogram labeled "RAFOS" is the signature of a swept-frequency signal (a "chirp") from one of the acoustic beacons that make up a sort of underwater GPS system for the eastern Pacific Ocean. The signal shown is from a source referred to as V1, moored off shore of Portland, Oregon at the approximate sound channel depth. The delay between the know broadcast time and detection time can be translated into a distance from the source. The signals from the five active sources in the eastern Pacific permit a drifting receiver to determine its position to an accuracy of about one km. Plotting the position of each drifting instrument allows determination of the eastern Pacific subsurface ocean currents, something otherwise very difficult to measure.
This positioning system is used by a group from the Monterey Naval Postgraduate School (NPS) to track "Lagrangian drifters" mapping subsurface currents in the eastern Pacific. The Pioneer Seamount Observatory plays a supporting role by monitoring the arrivals from each source, to keep track of which sources are active and to check their timing accuracy. The propagation time over these known paths through the ocean can be used to determine the sound speed, and this data is when extended over longer time intervals is expected to provide information on the ocean environment.
Over shorter time scales, variation in the arrival times of these signals can give an indirect measurement of subsurface currents. This measurement depends on the motion of the source in response to the current, a motion which sailors refer to as the watch circle, where a anchored boat drifts from one side of its anchor point to another as the current direction changes. The RAFOS sources are moored one to two kilometers above the ocean floor, at a depth of about 600m. At this depth, the motion of the water is dominated by cyclonic gyres called mesoscale eddies, or "middies." This motion translates into an observable variation in the travel time for the RAFOS signals.
Earthquakes and LFA. The signal from a small earthquake is indicated on the spectrogram. Such quakes are detected about once per day. While their power spectrum peaks at a few hertz, below the PSM cutoff of 6 to 8 Hz, their higher-frequency components are easily recognized in the spectrogram. These arrivals will eventually be integrated with seismometer data to study earthquakes in the Pacific floor, although at present there are no ocean-floor seismometers in this general area of the Pacific. At present, study of plate-tectonic motion along the California coast is hampered by the fact that most observations are made east of the plate boundary. The addition of a seismometer would be very valuable for earthquake geologists.
The recent announcement by the US Navy of its intention to test the SURTASS low-frequency active (LFA) sonar system for submarine detection in the Pacific lends additional interest to underwater acoustic monitoring. The frequency range (100-300 Hz) of the LFA source is well within the 10-450 Hz capability of PSM. The proposed source level of the LFA array is 240 dB re 1 mPa at 1 m. With the array operating 200 miles off the California coast, sound levels of 180 dB re 1 mPa would be observed at Pioneer Seamount and in the MBNMS. This is a sound level considered by some to be dangerous to marine mammals. Independent monitoring of these sounds at PSM during these tests would enable the sanctuary to quantify the noise levels produced and perhaps to look for effects on the behavior of marine mammals.
The Future of the Pioneer Seamount Observatory. Pioneer Seamount went off the air on September 24, 2002 at 12:07 UT. The center conductor of the coaxial cable is apparently shorted to sea water. As explained on the PMEL web site, "Testing by University of Washington engineers indicates a cable fault approximately 25 miles offshore. A decision will be made whether to repair the cable or not. No data will be available until the cable is repaired." While the point of failure appears to be in relatively shallow water, between 200 and 500 m in depth, the cost of deploying a ship to bring the cable to the surface and repair it would probably be over one hundred thousand dollars. NOAA officials will have to decide whether maintaining such a facility warrants this expense.
The first year of operation of this cable has revealed the variety and quality of information to be obtained from a cabled off-shore acoustic observatory. Continued operation of this facility should enhance the Sanctuary's ability to track populations of marine animals and to monitor their condition over a long time period. Measures of activity such as the rate of visitation of the Pioneer Seamount site by blue whales can be determined in an objective, bias-free way and with a minimum of scientist and ship effort. And with the use of air guns for oil exploration and the prospect of LVA operation nearby, continuous monitoring of noise in the sound channel is becoming more and more important. This observatory is unique in the world in providing real-time public access to a deep-ocean acoustic observatory.
A logical future upgrade to the PSM instrumentation package would include a horizontal hydrophone array, to give the direction of sound sources, and a seismometer. With these enhancements this facility could provide the public with a unique window into Sanctuary waters for many years to come.