Australian Integrated Marine Observing System (IMOS) Acoustic Observatories 1

Transcription

1 Australian Integrated Marine Observing System (IMOS) Acoustic Observatories 1 Summary The Acoustic Observatories sub facility archives ocean noise from around Australia at three primary locations: 32 o S off the east and west Australian coasts (since 2008 in Western Australia and 2010 in eastern Australia); and south of Portland, Victoria (since 2009) and makes the data publicly available. The facility has been paused in pending Australian Commonwealth Government funding decisions. The moorings are, or were located near or on, the continental shelf break. A passive acoustic mooring comprises a calibrated (2 Hz to 2.5 khz) autonomous sea noise logger set on the seabed recording 5 minute ocean noise samples at 6 khz sample rate every 15 minutes for months. In several instances grids of three or four instruments have been set up for tracking purposes, although more recent deployments use one or two instruments only. Currently we have a fourth site being sampled, on the shelf edge west of Kangaroo Island in South Australia. The Western Australian Government funded three years of data collection ( ) at 15.5 o S and 19.5 o S on the shelf edge of the northern WA coast. The passive acoustic observatories record sound emitted by natural processes in the ocean, underwater noise sources of biological origin, such as marine mammals, crustaceans or fish, plus man made noise sources. Through analysis of these signals it is possible to discriminate and identify different animal species and to assess the relative number of animals present within the range of acoustic observation, which can then be linked to ocean productivity or yearly migratory passage for great whales. Sea noise can be visualised on the IMOS web page, small amounts downloaded or large amounts requested from IMOS. Data is currently supplied in its raw format only, although we are considering making available processed detection times of various whales for selected sites. Please note that if requesting data from IMOS the best format to ask for is the raw data (*.DAT files), IMOS will offer *.wav file versions of the data but without a list of magic numbers the *.wav data is uncalibrated, whereas the raw data is calibrated. Instrumentation and Data The passive acoustic observatories comprise a sea noise logger placed on the ocean floor attached to a mooring which on command, releases floats to the surface. The sea noise loggers were designed and built at Curtin University and are well proven fully calibrated instruments (see McCauley et al. 2017). Each mooring also has ancillary temperature loggers attached with one on the seafloor and one between m above the seafloor. Sea noise data sets can be viewed via the AODN web portal: as time stacked spectrograms (~ 15 days, x scale) on a logarithmic frequency scale (y scale, 5 Hz to 3 khz) with intensity colour coded. Choose a deployment then scroll through in time. By left clicking a location on the spectrogram an image of the nearest sample's spectrogram and its waveform are displayed (while this feature should work, it may not work for some browser types). The time stacked spectrograms are useful for easily visualising when major sources are present as the dominant sources have a unique frequency content, so fill certain bands of energy in the spectrograms. For example a section from the Perth Canyon 1 Provided by Dr. Robert McCauley, Curtin University, Australia 21 February 2018 version 1

2 with nearby pygmy blue whale calling is shown on the figure below, with a full pygmy blue whale song type on the left and two days of stacked sea noise on the right. Figure showing on the left a pygmy blue whale three part song and on the right two days of sea noise containing thousands of pygmy blue whale calls. The frequency alignment can be easily seen here. A vessel passes during the middle time period. The archived sea noise contains information on whatever noise sources were active at the time, within the listening range of the receiver. What the listening range is for the receiver depends on a variety of factors, including the source type, its intensity, location in the water column, combined with the sound transmission environment along the path to receiver and the local ambient noise field. Typical detection ranges for the locations sampled are km for blue whales in the deep ocean and perhaps km back onto the continental shelf, or km for humpbacks on the shelf. Processing sea noise is not an easy task and to do it properly requires some degree of understanding of underwater acoustics and signal processing. There are tools developed for this (see links below) and to simplify interpretations we are intending to load onto the AODN portal time stamps of detection algorithm outputs with identification of when different whale types occurred, initially along the eastern Australian coast using the NSW receiver data. The archived sea noise contains an invaluable record of the vocalising and physical noise sources. Applications of Data Examples of applications of the data include: An air craft crashing into the ocean may generate underwater sound by its impact at the surface or by debris imploding as it sinks. A signal was detected on two IMOS receivers coincident with the best estimate of the time of the loss of Malaysian aircraft MH370. After identification on the Perth Canyon IMOS receiver this signal was found in the Comprehensive Test Ban Treaty Organisation (CTBTO) hydrophone sensors at the Cape Leeuwin (HA01) site which gave a bearing consistent with the believed crash area. The signal falls almost precisely at the believed crash time but is not located along the search path given by analysis of the aircraft satellite communications (so called "7th arc"). Using data from the IMOS Scott Reef, IMOS Perth Canyon and CTBTO hydro acoustic stations the source location is some 8000 km from the Perth Canyon, slightly west of the Maldives assuming the same source was received at each site. The signal was later found on the CTBTO Diego Garcia hydrophones where it had arrived by horizontally refracting around the Diego Garcia Island group. This is a good example of long range ducting of signals in the open ocean due to the "deep sound channel" created 2

3 by a minimum sound speed at around 1000 m depth in the deep ocean which traps sound energy over a narrow frequency band (5 100 Hz) by refraction. Sound energy over this frequency band suffers practically no absorption losses so can travel ocean scale distances losing energy only by 2 dimensional spreading. Currently we believe this signal was generated by a small earthquake and was not related to the loss of MH370 although no matching earth tremor signals have yet been identified on the earthquake seismic network. References are Duncan et al. (2014 a & b). An example of the types of sources commonly found in the IMOS passive acoustic observatories was published in Erbe et al. (2015). Using four years of data this paper summarises the major contributors to sea noise in the Perth Canyon, these being whales, fish, wind, rain and ships. Statistics of the sea noise are given along with examples of the contribution of these sources to sea noise, their spectral characteristics, temporal patterns and how the sea noise can be used to define physical and biological parameters. A comprehensive analysis of fish chorusing behaviour from the Perth Canyon has been carried out, with the aim of identifying the chorus source (McCauley and Cato, 2016). Fish chorusing behaviour in the oceans is common but hardly reported or known of in mainstream biological sciences. The paper by Erbe et al. (2015) mentioned above presents a summary of the Perth Canyon fish choruses while the McCauley and Cato (2016) paper gives details of this chorus and suggest the most likely source is fish of the family Myctohidae. The detailed analysis suggest the chorus, which encompasses spatial scales in excess of the tens of nautical miles sampling program carried out and which occurs nightly and year round, is associated with the rise of the deep scattering layer and is found in regions of known highest secondary productivity in the Perth Canyon. This chorus is currently believed to be produced by small fishes foraging high in the water column. A figure showing the trends in this chorus over 4.3 years of IMOS data and its regular, nightly and seasonal occurrence is shown below. Figure showing the trend in fish chorus level activity in the Perth Canyon over a 4.3 year period using IMOS data. The top plot shows the level across an evening in the fish chorus frequency band stacked over 1,574 days with each evenings times zeroed to time of local sunset and intensity as the colour scale. The heavy black line is the time of sunrise. The lower plot shows the level in the chorus frequency band each evening over 0.4 to 5 hours post sunset. Several PhD students are making use of IMOS data. This includes further work on fish chorusing patterns, fin whale patterns around Australia or studies onto how sea noise derived indexes of whale calling can be turned into density estimates of the number of whales present in the listening range of 3

4 the sea noise logger. These are works in progress. The 2014 Perth Canyon IMOS sea noise data is being used by one of the students to correlate against a high resolution tag attached to a pygmy blue whale by the Centre for Whale Research (WA), which gave eight days of fine scale whale diving behaviour. The tag data suggest the whale was a singer, thus we are attempting to correlate pygmy blue whale song from the IMOS receiver with the tag. Useful links If you have any questions regarding the data, or corrections, or would like to add a publication or presentation that uses IMOS data please contact the IMOS office via publication(at)emii.org.au. Contact(s): A. Prof Rob McCauley r.mccauley@cmst.curtin.edu.au or A. Prof. Alexander (Sascha) Gavrliov, a.gavrilov@cmst.curtin.edu.au Your access to Acoustic Observatory data discovery and exploration is through the IMOS Ocean Portal: Recommended software packages for processing data are indicated below. If you wish to use either package and it cannot open the IMOS data files please advise us and we will contact the software providers to advise them of the data format. We will provide routines to read the data files if requested. IMOS data has been analysed using Ishmael. Curtin University has a processing package used to easily and quickly assess these data sets (CHORUS, which does the same as the web viewer but with more options) which is available as below. Ishmael Pamguard CHORUS software/ Deployments A map of the sites sampled is shown below. Sites sampled up until are located: off the central NSW coast working out of Tuncurry / Forster (32 o 16.2' S, 152 o 57.0' E); south of Portland, Victoria (38 o 32.5' S, 115 o 0.1' E) with Portland the home port; west of Kangaroo island (36 o 7.6' S, 135 o 55.0' E) with this site serviced by the South Australian Research and Development Institute on the vessel RV Ngerin; and west of Perth in the Perth Canyon (31 o 52.0' S, 115 o 0.1' E). In addition to this two sites were sampled from with support from the Western Australian Government, these sites were located at: "Dampier" (19 o 23.3' S 115 o 54.9' E); and "Scott Reef" (15 o 29.0' S, 121 o 15.1' E) with the site names indicative only. The locations of all sites are shown below. The Perth Canyon has been sampled since 2008, Portland since 2009, NSW since 2010, Kangaroo island since 2015 and the north western WA sites over Exact coordinates and times of each deployment are listed in the table below. Note due to funding cuts the IMOS passive acoustic facility has been paused post gear recovery in We are hoping that we can re deploy in financial year 2019/2020. This will only happen if potential users can pressure IMOS into maintaining the data stream. 4

5 Locations of sites currently sampled for sea noise by IMOS (black) and the North Western Australia locations (red) which ceased operation in mid Sites were chosen to be on or near the continental shelf break edge in order to access signals transmitted in the deep ocean. All sites are either on the shelf break (Perth Canyon) or within a few nautical miles inside of the shelf edge as given by the 200 m depth contour. The Perth Canyon site was chosen as it was known to be a 'hot spot' for several species of offshore great whales and as data was available here from 2000 onwards (held by Curtin University). The NSW site was chosen for multiple reasons, as its at the same latitude but on the opposite side of Australia to the Perth Canyon, its an area which is completely undersampled from the perspective of long term sea noise records, as there are known to be regular, undescribed movements of great whales offshore up and down the coast, as we had a good fishing port with high quality fisherman close by and as it is relatively free of deep water trawling. We were initially hoping to document movement of the eastern Indian Ocean pygmy blue whale, which is common in western and southern Australian sea noise records, as it moved up and down the eastern Australian coast. As it transpired, this does not happen, rather based on the IMOS passive acoustic records there is a separate pygmy blue whale sub population which transits the eastern Australian coast, commonly referred to as the NZ pygmy blue whale but which would be more accurately termed the western Pacific pygmy blue whale. The Portland site was chosen as several years of sea noise were available from this site thus the shore based facilities were known to us, as its on what is termed the Bonney coast which has some of the largest seasonal upwelling events in Australia, the site is on a rocky ridge so is free of trawling and as its known to be commonly visited by various poorly known offshore great whale species. The Portland IMOS passive acoustic mooring is the only long term oceanographic mooring between the IMOS moorings in South Australia and moorings around Tasmania. The Kangaroo Island site was chosen as its near deep water IMOS oceanographic moorings thus making servicing easier and its in a highly productive but poorly known (from the perspective of whale presence) and sensitive area. The passive acoustic moorings are typically serviced every months, with the duty cycle giving an month duration (depending on batteries). We have not always been able to maintain this servicing frequency, notably there have been periods of funding uncertainty during which we did not know if we should re deploy again, so we had to wait before scheduling field work for re deployments. The NSW site is plagued by strong currents produced by the East Australian Current, which have precluded us from recovering gear, for almost a year in one instance. The moorings are designed to decouple the sea noise 5

6 logger and its hydrophone from the mooring, so reducing noise artefacts which are extremely easy to produce. Traditional in line oceanographic moorings do not work for passive acoustic instruments especially in areas of high current. Given the mooring design we cannot release our moorings in currents of greater than around 2 knots as the buoys get pulled down and it is highly dangerous and near impossible to grapple in these conditions. To date (2018) we have lost two moorings, one in northern Western Australia (Dampier, ) and one in the Perth Canyon. On recovery the instruments are calibrated post deployment (a gain with frequency check is made to compare against a similar calibration made pre deployment), the instrument clocks checked for drift against GPS, UTC transmitted time and the data copied across and backed up. We generally run a routine which then reads all data files and saves a master table for that deployment of file names and sampling times. The data is then viewed as for the IMOS web viewer to identify any notable events and to assess data quality. The Perth Canyon data is subject to Defence restrictions and we may have to remove sections for security reasons. In general Defence are aware of and try to avoid our equipment. Data is copied to a hard disk and sent to IMOS in Hobart, whom archive it and make the data available upon request. Instrumentation The sea noise loggers used were designed at Curtin University in the early 2000's and are well proven instruments (see McCauley et al. 2017). The housing (stainless, 1.2 m long x 115 mm dia., ~ 34 kg in air) lies on the seabed in a frame with an external, weighted hydrophone (Massa TR1025C or HiTec U90 types). A pre amplifier (20 db gain) inputs to a signal conditioning card where a low frequency rolloff is applied to frequencies below 8 Hz to flatten the naturally high levels of low frequency sea noise and increase the system dynamic range. The signal is digitised (16 bit), a further 20 db of gain applied and samples saved to a small capacity flash card in PC format. When the small flash card is around 3/4 full files are copied to either a 128 GB hard drive or flash card (128 GB set by the 32 bit operating system). We can rack the larger storage drives together but have opted to use one only to keep power consumption reasonable. The systems were designed to sample intermittently and to use small flash cards initially, in order to reduce power consumption (cycled alkaline batteries last considerably longer than those continuously drained), system self noise (HDD, some flash cards and some SD cards can all generate high electronic noise) and time lags inherent in reading the FAT tables of large capacity drives. The moorings are designed to be deployed from commercial fishing vessels hence use rope, and to isolate the sea noise logger as best as possible from the mooring to reduce noise artefacts. A mooring comprises the sea noise logger, a ground line around twice the water depth laid on the seabed with attached weights, floats and an anchor, which couples into a swivel 5 m above an acoustic release, with 175 kg of dump weights below this and a series of floats above the swivel. An image of a mooring is displayed below. These moorings are routinely deployed to 450 m depth. We have had some peculiarities with the moorings, in the NSW and southern Australia moorings leatherjacket fishes eat the ropes and hydrophone cable, we have to armour the hydrophone cable and have to use a hard lay thick rope. We prefer to liaise with professional fisherman, they offer a local base, a wealth of local knowledge, can do anything at sea and enjoy learning about their environment. An image of two sea noise loggers and the staff involved with gear recovery and deployment in 2013 out of Portland is shown below. The sea noise recorders are calibrated for system gain with frequency by inputting white noise of known level (always 90 db re V 2 /Hz) in series with the hydrophone. This calibrates the full system response (> 1 2 Hz) and accounts for the impedance matches of various components. To calibrate the systems below 2 Hz we can inject the white noise in at 70 db re V 2 /Hz, although most of the hydrophones used will only sample down to around 1 Hz. We have deployed hydrophones capable of sampling down to near DC frequencies. The hydrophones used and the calibration technique give a calibrated response over 2 Hz up to the anti aliasing filter frequency used, 2.8 khz. The calibration files are available through IMOS on request. The instrument clocks are set to GPS transmitted UTC time before deployment and the clock drift 6

7 read after recovery using hardware and software. Due to often sharp jumps in temperature during deployment and recovery we quote an absolute clock accuracy of ± 250 ms. Metadata lists for each deployment are available from IMOS. These list location, sample times, calibration information (calibration file, hydrophone details and clock synchronisation) and notes, Sea noise files contain header and footer information in text format, with 16 bit unsigned integer data in binary format inserted between (0 5 V internal rail). The header contains information on sample set up and scheduling, the footer the time sampling started and ended (noting a 'tick' is 1/65536 of a second and is always added to the seconds displayed). Note that the time sampling started is not the same as the time the file was opened (encoded in the file names as seconds from 01 Jan 1970 in hexadecimal format). The file format is relatively simple and a Matlab file for reading raw data is available from IMOS. D ~ m Not to scale Ground line = 220 x 4 = 880 m Riser = 35 Total line = 915 m ~ 2:1 Weight along line = = 26 kg (5 kg for shackles, thimbles etc) Buoyancy on riser = (9 x 7.5) + (3 x 2.5) = 75 kg Buoyancy once released = 75 + (4 x 2.5) - 26 = 59 kg 3 x 200 mm buoys ~ 50 m 10 m 175 kg Acoustic Release/s 3 x strings 3 x 280 mm buoys 200 mm Ø buoy 220 m 220 m 5-7 kg 5-7 kg ~ 750 m 200 mm Ø buoy 5-7 kg 220 m m 30 m Noise logger 5 1 m Figure displaying a schematic diagram of a sea noise logger mooring. All lines are 14 mm hard lay polypropylene rope. The mooring ground line length is scaled for the water depth. 7

8 Image of staff associated with sea noise logger work out of Portland, Victoria with two recently recovered sea noise loggers (had been down for 12 months) on the table. Rob McCauley, the passive acoustic observatory manager is on the right, three fisherman in the centre and a field assistant on the left. 8

9 Image of a sea noise logger being recovered off Portland. 9

10 Data Your access to Acoustic Observatory data discovery and exploration is through the IMOS Ocean Portal [link] A summary of IMOS passive acoustic data collected up to with one representative instrument for each deployment period/site highlighted is given below in Table 1. The set number is a number allocated by Curtin for that deployment and is not necessarily in order of time or location. Note that temperature data is available from each site for each deployment, contact IMOS or the passive acoustic facility manager (McCauley) for this data. Note in many deployments multiple loggers were set per deployment and for tracking grids each logger had two sample schedules, the normal 5 minute 6 khz samples every 15 minutes and once per day a shorter 20 or 22 khz sample. The instances where multiple loggers were used per deployment have been for tracking purposes, the loggers were set in a grid of four, 1) to the SE, 2) to the SW, 3) to the north in an equilateral triangle of approximately 5 km sides, and 4) in the triangle centre. An acoustic release on the centre mooring was set to ping over 35 minutes every 20 s at 7.5 khz. These pings were captured by the 22 khz samples (also by the 6 khz samples) and used to synchronise the instrument clocks (set three clocks relative to a fourth). Analysis of these data sets is complex, especially for synchronising the clocks, so users are asked to contact Rob McCauley ( below) if they wish to pursue tracking analysis. See Gavrilov et al. (2012) for an example of the tracking capability. The data can be used for various purposes including (but certainly not limited to): studies of vessel traffic Passing vessels produce distinctive noise signatures, these can be related to vessel movements; recording man made noise such as vessel or marine seismic noise Moored sea noise loggers can give long term statistics on the amounts of noise produced by various man made activities, with two of the more common noise sources being ships and offshore marine seismic surveys; studies into ambient noise Wind, rain, ice related noise, earthquakes even swell and sea state are some examples of relatively common natural sources which greatly influence ambient noise levels. The ambient noise statistics can be used in turn, to define the physical factors raising sea noise, such as wind speed or rainfall frequency and intensity. studies into fish presence Fish can call independently, commonly en masse in large schools, or somewhere in between, often regularly raising ambient noise levels in the frequency band of their calling by significant amounts. By knowing the characteristics of a fish species call type (they generally have only a limited repertoire) and having some knowledge of why and how they are producing the call, then one can obtain long term information on trends and behaviour for different fish species. Since a sea noise logger is integrating information over a large area, approximately one km for an individual fish call or perhaps km for fish choruses as detected by the IMOS observatories, then the sea noise offers a long term, large scale technique for monitoring. studies into whale habits, migrations and seasonal presence Most baleen whale species produce vocalisations which may be detected at many tens of km. The IMOS sea noise loggers are optimised to detect these signal types. The sea noise data allows yearly monitoring of these whales, for example allowing us to define which species will travel up and down our coasts, when they may go past and with work, how many pass each year. This technique is the only reliable way to monitor our offshore great whales, which the IMOS sea noise loggers have shown are common and involve a multitude of species from all the primary sites. One of the great powers of the sea noise logger sampling is the high time definition available over long time frames. This allows the data to be used to study natural rhythms in calling animals at a resolution which is 10

11 simply not available to any other technique in the ocean. Marine animals do not have watches and have never heard of the Julian calendar, the precursor to our current Gregorian calendar. Their rhythms are based on day length, moon phase and summer and winter solstices in complex ways that the sea noise logger data is beginning to elucidate. An example of the seasonal nature of fish calling is presented in McCauley et al. (2012). As an example of what the IMOS sea noise loggers can tell us in the long term, consider pygmy blue whales in the Perth Canyon. We have been monitoring these whales with IMOS gear since 2008 and in total since We can now reliably predict the seasonal visitation of these whales each year, using data from previous years adjusted for the whale's seasonal clock which is not based on our Gregorian calendar, but displayed below on our Gregorian calendar where the predicted pygmy blue whale visitation for 2016 is shown. The pygmy blue whales arrive in the Perth Canyon on their northbound migratory leg, will stay if there is food present, then depart. Using each of these stages: 1) arrive; 2) period when they may stay to feed; and 3) migratory tail, and the corrected seasonal time base, we can look at the trend across time of the population based on numbers of whales calling. This is also shown as a figure below. Figure showing predicted occurrence of pygmy blue whales in the Perth Canyon based on previous data. The right hand scale is the mean number of calling pygmy blue whales per day. 11

12 Figure showing the trends in pygmy blue whale calling (mean number of calling individual whales per day) for the early north bound migratory period (red circles), the period when the may stay and feed (black squares) and the tail of the migratory leg (magenta triangles). Staff The passive acoustic sub facility is operated by the Centre for Marine Science and Technology of Curtin University. The primary staff associated with the project are listed below with staff shown on an image below. Robert McCauley (sub facility manager, field science leader) primary contact: Alexander Gavrilov (data analysis), second contact: Christine Erbe (Director, CMST) Mal Perry (technician, moorings) Dave Minchin (technician, instrument calibrations) References Duncan, A. J., Gavrilov, A., McCauley, R.D., (2014a) Analysis of Low Frequency Underwater Acoustic Signals Possibly Related to the Loss of Malaysian Airlines Flight MH370. Prepared for: Australian Transport Safety Bureau. Centre Marine Science and Technology Curtin University, R , Jun 2014 Duncan, A. J., McCauley, R.D., Gavrilov, A., (2014b) Results of analysis of Scott Reef IMOS underwater sound recorder data for the time of the disappearance of Malaysian Airlines Flight MH370 on 8th March Centre Marine Science and Technology Curtin University, Sep 2014; available from a.duncan@cmst.curtin.edu.au or r.mccauley@cmst.curtin.edu.au Erbe, C., Verma, A., McCauley, R. D., Gavrilov, A. Parnum, I. (2015) The marine soundscape of the Perth Canyon. ScienceDirect, 137:38 51 Gavrilov, A.N., McCauley R.D., Pattiaratchi C., Bondarenko, O. (2012) The use of passive acoustics to observe the presence and movement of pygmy blue whales (Balaenoptera musculus brevicauda) in the Perth Canyon, WA. Proceedings of the 11th European Conference on Underwater Acoustics, July 2012 Edinburgh, UK. Proceedings Institute Acoustics 34(3) McCauley RD., (2012) Fish choruses from the Kimberley, seasonal and lunar links as determined by long term sea noise monitoring. Australian Acoustical Society Proceedings of Acoustics 2012 Fremantle 6pp. McCauley, R. D., Cato, D. H. (2016) Evening fish choruses in the Perth Canyon and their potential link with Myctophidae fishes. J. Acoust Soc. Am. 140(4): McCauley, R.D., Thomas, F., Gavrilov, A. N., Duncan, A. J., Cato, D., Parsons, M., Parnum, I. M., Salgado Kent, C. Christine Erbe, C. (2017). A history of recording underwater sound around Australia and the Curtin Underwater Sound Recorder hardware: A researchers perspective. Acoustics Australia, 45(2), , DOI /s

13 Image of CMST staff with: Christine Erbe (front row, second from right); Alexander Gavrilov (back row left); Robert McCauley (far back right); Mal Perry (second row from back, centre); Dave Minchin (second row from front, far right). 13

14 Table 1: Details of all IMOS passive acoustic deployments which have been recovered with a single instrument indicative for each deployment highlighted in bold. The columns are: set the Curtin set number; latitude (all S); longitude (all E); start of sampling in water (UTC); end of sampling in water (UTC); days sampled; water depth (a zero implies its not entered); samples made while the instrument was in the water (samples are files sorted by date/time); sample length in seconds; time increment between consecutive samples (minutes); notes. The Notes include the instrument serial number (usually preceded by the E) and the sample rate. Set Latitude Longitude Start time End time days depth Good samples Length (s) Feb :00 21-Apr : (5200) Perth Canyon 2008 Silicon 6 khz Dec :55 02-Oct : (22065) Perth Canyon 2009 E04 6 khz Feb :45 12-Oct : (21549) Perth Canyon 2009 E03 6 khz Feb :00 02-Oct : (21471) Perth Canyon 2009 E02 6 khz Feb :15 02-Oct : (21887) Perth Canyon 2009 E07 6 khz Feb :08 01-Oct : (195) Perth Canyon 2009 E04 22 khz Feb :08 01-Oct : (197) Perth Canyon 2009 E02 22 khz Feb :09 01-Oct : (185) Perth Canyon 2009 E07 22 khz Feb :08 11-Oct : (194) Perth Canyon 2009 E03 22 khz Nov :15 23-Jul : (23984) Perth Canyon E06 6 khz Nov :15 23-Jul : (23878) Perth Canyon E05 6 khz Nov :15 22-Jul : (24084) Perth Canyon E02 6 khz Nov :45 23-Jul : (21364) Perth Canyon E03 6 khz Nov :07 21-Jul : (204) Perth Canyon E06 22 khz Nov :07 21-Jul : (218) Perth Canyon E02 22 khz Nov :07 22-Jul : (225) Perth Canyon E05 22 khz Dec :07 17-Mar : (94) Perth Canyon E03 22 khz Aug :15 10-May : (26378) Perth Canyon E06 6 khz Aug :15 11-May : (26625) Perth Canyon E05 6 khz Aug :15 09-May : (26234) Perth Canyon E02 6 khz Inc. (min) Notes 14

15 Aug :15 08-May : (26335) Perth Canyon E03 6 khz Aug :07 08-May : (258) Perth Canyon E03 22 khz Aug :07 10-May : (265) Perth Canyon E05 22 khz Aug :07 03-May : (255) Perth Canyon E02 22 khz Aug :07 09-May : (261) Perth Canyon E06 22 khz Jul :15 18-Jun : (32433) Perth Canyon E03 P3 6k Jul :15 17-Jun : (32490) Perth Canyon E02 P4 6k Jul :15 20-Jun : (32659) Perth Canyon E05 P1 6k Jul :05 17-Jun : (312) Perth Canyon E03 P3 22k Jul :05 16-Jun : (322) Perth Canyon E02 P4 22k Jul :05 18-Jun : (329) Perth Canyon E05 P1 22k Aug :15 14-Jun : (29854) Perth Canyon E09 6 khz Aug :15 14-Jun : (29898) Perth Canyon E05 6 khz Nov :00 31-Mar : (11801) IMOS PerthCanyon E Nov :05 04-Nov : (32637) Perth Canyon E24 6kHz Nov :50 31-Mar : (239) IMOS PerthCanyon E09 20 khz Nov :55 03-Nov : (635) Perth Canyon E24 20kHz May :15 22-Dec : (21959) Portland IMOS 2009 P3 6 khz E May :15 22-Dec : (21982) Portland IMOS 2009 P1 6 khz E May :45 04-Dec : (20062) Portland IMOS 2009 P2 6 khz E May :08 21-Dec : (184) Portland IMOS 2009 P1 22 khz E May :08 20-Dec : (187) Portland IMOS 2009 P3 22 khz E May :08 02-Dec : (175) Portland IMOS 2009 P2 22 khz E Feb :15 30-May : (10803) Portland 2010 P2 E12 6 khz Feb :15 25-Sep : (21904) Portland 2010 P1 E11 6 khz Feb :08 30-May : (86) Portland 2010 P2 E12 22 khz Feb :08 24-Sep : (184) Portland 2010 P1 E11 22 khz 15

16 Dec :15 03-Dec : (32227) May :00 26-Nov : (19332) Portland P4 E27 6k Portland P3 E26 6k Dec :15 05-Dec : (34531) Portland P1 E34 6k Feb :00 06-Nov : (25198) Portland E21 6k Feb :15 06-Nov : (25202) Portland E22 6k Feb :51 05-Nov : (253) Portland E21 6k Feb :56 05-Nov : (247) Portland E22 20k Jul :45 25-Sep : (6976) Portland IMOS 2009 P4 6 khz E Jul :08 24-Sep : (51) Portland IMOS 2009 P4 22 khz E Nov :00 17-May : (18348) Portland P2 E08 6 khz Nov :00 17-May : (18327) Portland P1 E11 6 khz Dec :30 27-Nov : (31844) Dec :05 27-Nov : (31852) Portland 2014 E Portland 2014 E Feb :00 04-Oct : (22045) NSW 2010 P1 E8 6kHz Feb :00 28-Sep : (21892) NSW 2010 P4 E7 6kHz Feb :00 28-Sep : (21984) NSW 2010 P3 E4 6kHz Feb :53 04-Oct : (232) NSW 2010 P1 E8 22kHz Feb :54 27-Sep : (205) NSW 2010 P4 E7 22kHz Feb :53 27-Sep : (187) NSW 2010 P3 E4 22kHz Apr :15 25-Apr : (36823) NSW E07 P3 6k Apr :15 26-Apr : (36968) NSW E11 P1 6k Apr :15 27-Apr : (37137) NSW E04 P2 6k Apr :04 24-Apr : (373) NSW E11 P1 22k Apr :04 27-Apr : (381) NSW E04 P2 22k Apr :04 25-Apr : (374) NSW E07 P3 22k 16

17 Jun :15 30-May : (34249) NSW E27 6 khz Jun :15 30-Apr : (31268) NSW E12 6 khz Jun :15 27-May : (33515) NSW E26 6 khz Jun :15 01-Jun : (34543) NSW E34 6 khz Jun :05 01-Jun : (339) NSW E34 22 khz Jun :05 28-May : (316) NSW E27 22 khz Jun :05 29-Apr : (304) NSW E12 22 khz Jun :05 26-May : (313) NSW E26 22 khz Nov :00 29-Sep : (30042) Scott Reef 6 khz E Oct :00 02-Jun : (23410) Scott Reef khz E Aug :39 08-May : (25061) Scott Reef E Nov :00 27-Sep : (29848) Dampier 6 khz E Aug :30 13-Jun : (29369) Dampier E07 17