The frigate HMCS Ottawa (credit: marinetraffic.com)

Mid-frequency active (MFA) sonar pings were detected this morning (Monday, 2/6/12) on the Salish Sea Hydrophone Network by Jeanne Hyde and signal processing computers. Utilizing online ship tracking data from the Automatic Identification System, Jeanne found the Canadian Naval vessel HMCS Ottawa (MMSI 316195000) entering Haro Strait as she continued to hear “pings” from its SQS 510 (V) mid-frequency active sonar on the live hydrophone located at Lime Kiln State Park. Her experience is detailed on her blog, Whale of a Porpoise.

Here’s an example of what she heard: a single ping from the SQS 510 sonar.

Jeanne’s recording (below) begins at 04:49 and ends at 05:39.  82 pings occur in the first 38.25 minutes of the recording, resulting in an average ping rate of 1 ping every 28.35 seconds.  Extrapolating back to the first automated detection of a ping (at Lime Kiln) at 04:42:50, there were likely at least 14 pings emitted before Jeanne started recording.  So, the minimum total number of pings emitted is 96.

41-minute recording of 82 pings (15 Mb mp3 file) by Jeanne Hyde from the Lime Kiln live hydrophone stream

This is the first detection of Canadian MFA sonar on our hydrophone network, and — to our knowledge — the first time Canadian sonar has been used in the U.S. critical habitat of the Southern Resident killer whales, an iconic marine mammal population that is listed as endangered on both sides of the border. While policies guiding Canadian Naval use of mid-frequency sonar in U.S. critical habitat are not clear to us, there is guidance on the web regarding Canadian use of Canadian Low Frequency Active (LFA) Sonar (at least in the Atlantic).

SQS 510 sonar on a Halifax hull (credit: hazegray.org)

As a start to evaluating the potential acoustic impacts of this incident, we report here on the preliminary hydrophone and vessel location data collected by the Network during this incident, including: human detections and recordings; automated detections and recordings; and AIS data logged at the Lime Kiln. We are also aggregating information from regional sighting and listening networks about species that may have been ensonified. As additional data and analyses become available, updates will be posted here. Authors of this analysis are bioacoustians: Jason Wood, Research Associate with The Whale Museum; and Scott Veirs, of Beam Reach Marine Science and Sustainability School.

Below are the compressed (mp3) recordings and coarse spectrograms of the sounds that were auto-detected this morning. They begin with a series of low frequency sounds and echoes that may have been from an impulsive source, like a detonation or explosion. Then the series of high-frequency pings occurs between 4:42:50 and 5:08:17 at three network locations: Lime Kiln (13 pings), Port Townsend (1), and Orcasound (1). While we are not yet sure if pings were detected at Neah Bay or on the NEPTUNE Canada hydrophones located near the entrance to the Strait of Juan de Fuca, it appears that the sonar ensonified a good portion of the eastern Strait of Juan de Fuca and southern Haro Strait.

First there were 4 detections of explosive sounds (3/12/2012 update: 5 detections)

Note of clarification added on 2/9/12 at 13:30: In the following table of audio recording files and spectrogram images, the “node” or hydrophone location is denoted with “lk” meaning Lime Kiln State Park, “os” meaning Orcasound (5 km north of Lime Kiln), and “pt” meaning Port Townsend Marine Science Center. The “dB” label offers an estimate of the 30-second average (RMS) received sound pressure level at the time the recording was triggered, and therefore should not be taken as an indication of the receive level of any particular sound heard within the recording (i.e. an explosion or sonar ping) The trigger label indicates whether the sound was recorded because it was exceptionally loud (“PWR”) or because it had the changing pitch and duration typical of a killer whale call (“PKT”).

3/12/2012 update — 1 Detection at Orcasound (~5km north of Lime Kiln)

2012-02-06 04:31:05 node=os dB=107 trigger=PWR

3/12/2012 update — The Lime Kiln recordings were made from a single Reson TC4032 (on both channels in the stereo recordings). The single detection at Orcasound does not match the arrival times of auto-detected explosive-sounds at Lime Kiln within the few seconds that would warrant estimation of the bearing to the source (or at least suggest whether the source was north or south of the hydrophones. The only explanation for the distinct arrival times are: (1) the sound source was too faint to be detected at both locations; (2) the computer clocks were not synchronized to NIST (though software was installed to do so at all hydrophone network locations); or (3) (more likely) something caused the longer-term mean receive level to rise at LK or OS to the point that the PWR trigger threshold was exceeded only at the quieter location when a given explosive-like sound arrived.

2012-02-06 04:32:37 node=lk dB=102 trigger=PWR	2012-02-06 04:33:08 node=lk dB=103 trigger=PWR
2012-02-06 04:38:35 node=lk dB=103 trigger=PWR	2012-02-06 04:39:07 node=lk dB=104 trigger=PWR

Then there were 14 detections of sonar pings

Note of clarification added on 2/9/12 at 13:30: In the following table of audio recording files and spectrogram images, the “node” or hydrophone location is denoted with “lk” meaning Lime Kiln State Park, “os” meaning Orcasound (5 km north of Lime Kiln), and “pt” meaning Port Townsend Marine Science Center. The “dB” label offers an estimate of the 30-second average (RMS) received sound pressure level at the time the recording was triggered, and therefore should not be taken as an indication of the receive level of any particular sound heard within the recording (i.e. an explosion or sonar ping) The trigger label indicates whether the sound was recorded because it was exceptionally loud (“PWR”) or because it had the changing pitch and duration typical of a killer whale call (“PKT”).

2012-02-06 04:42:50 node=lk dB=118 trigger=PWR	2012-02-06 04:43:44 node=lk dB=121 trigger=PWR
2012-02-06 04:44:12 node=lk dB=120 trigger=PWR	False positive deleted…
2012-02-06 04:44:41 node=lk dB=118 trigger=PWR	2012-02-06 04:45:35 node=lk dB=121 trigger=PWR
2012-02-06 04:46:32 node=lk dB=120 trigger=PWR	2012-02-06 04:47:00 node=lk dB=121 trigger=PWR
2012-02-06 04:48:01 node=pt dB=91 trigger=PKT	2012-02-06 04:48:52 node=lk dB=119 trigger=PWR
2012-02-06 04:49:22 node=lk dB=120 trigger=PWR	2012-02-06 04:57:50 node=os dB=126 trigger=PKT
2012-02-06 04:57:55 node=lk dB=106 trigger=PKT	2012-02-06 05:05:56 node=lk dB=114 trigger=PKT
2012-02-06 05:08:17 node=lk dB=115 trigger=PKT

While it is possible that other vessels were involved in the sounds recorded this morning, the location and public description of the Ottawa are consistent with what was heard. We are particularly curious about the source of the explosive sounds heard before the sonar pings.

According to hazegray.org, the Ottawa is a multi-role patrol frigate (FFH [Cardinal Mark N; Eog BEOFOR, "Ocean Beaver"?]) Halifax class vessel, 134 m long and X 16 m wide, displacing 4,470 metric tons, with maximum speed of 29+ knots and a compliment of 225 sailors. It carries two sonar systems: the hull-mounted mid-frequency SQS-510 and the low-frequency SQR-19(V) w/CANTASS towed array.

AIS systems report that it was last in port in San Diego, has been operating in the north Pacific, and was bound for Esquimalt mid-morning today. The “show track” feature of marinetraffic.com revealed mid-day today this track from 21:15 2/5/12 through 08:52 this morning (2/6/12).

Track of the Ottawa into the Salish Sea

Ottawa's track in critical habitat

A rough chronology from combining acoustic detections and AIS data (Google spreadsheet embedded below) suggests the following sequence of events. After steaming through the night through the Strait of Juan de Fuca, the Ottawa slowed down near Race Rocks and turned into the waters south of Victoria. It then reacted — possibly to detonation-like sounds that were auto-recorded at Lime Kiln — and then changed speed and course, began pinging, crossed into U.S. waters, turned northward at near maximum speed, stopped pinging and slowed down as it crossed over Middle Bank, proceeded further north, and then turned back to return to the Canadian waters south of it’s home port of Esquimalt.

Ottawa's track with estimated sonar start and stop locations.

A 50 millisecond sample of the starting frequencies of an example ping had peak power at 7042 Hz with a secondary peaks at 1523 Hz, 12615 Hz, and 21171 Hz (see figure below from Audacity). The peak power frequency is consistent with the 2-8 kHz frequency range specified for the SQS 510 sonar system, which is manufactured by General Dynamics Canada. Each ping had high intensity receive levels for ~0.5 second duration and pings were separated by about one minute.

Spectrogram of a typical ping.

Further details regarding the expected source and receive levels from the SQS 510 can be derived from its 1997 description in the book “The Naval Institute guide to world naval weapons systems, 1997-1998.” While it is likely that the system has been up-dated since 1997, this general description provides helpful characteristics:

SQS-505/509510/Type 5051

SQS-505 is the standard Canadian surface-ship sonar. Its 36 stave [10 elements/stave] cylindrical transducer (4ft dia x 4 ft) produces 35 10×10-deg beams at 6.4, 7.2, or 8.0 kHz [bandwidth 400 Hz divided into sixteen 25-Hz channels for each beam]. Modes: ODT, triple RDT (TRDT), directional transmission (DT), ASPECT, and swept FM (wide-band mode). Pulse lengths: 2.5, 10, 40, 480 msec. Source level: 126 dB (ODT), 135 dB (TRDT, DT). Signal processing: matched filter with time, function automatic gain control (TFAGC) or incoherent (wide-band) processing. Nominal range is 200-32,000 yd (range scales 2, 4, 8, 16, 32 kyd). Target Doppler can be up to 40 kt. Passive operation: 10 kHz. Track capacity is five targets.
….
The (V)6 version for the “City” class has new transmitter and receiver (source level 133 dB, pulse length 10 or 40 msec). Active frequencies match those of earlier versions, but passive frequency is reduced to 4.5 kHz and there is a steerable passive mode.
….

Our next steps are to estimate calibrated receive levels during the event and to gather more information about animals that may have been exposed to harmful sound levels. We may also attempt to estimate the approximate location of the explosive sounds.

In the interim, please report any observations of injured or stranded marine mammals in U.S. waters to NOAA by calling their hotline: 1-800-853-1964. Report unusual behavior or strandings in Candian waters to DFO by calling their hotline: 1-800-465-4336.