Search for Malaysia Airlines Flight 370

The Malaysian government mobilised its civil aviation department, air force, navy, and Maritime Enforcement Agency; and requested international assistance under Five Power Defence Arrangements provisions and from neighbouring states. Various nations mounted a search and rescue mission in the region's waters.[42][43] Within two days, the countries had already dispatched more than 34 aircraft and 40 ships to the area.[24][28][44]

On 11 March, the China Meteorological Administration[45] activated the International Charter on Space and Major Disasters, a 15-member organization whose purpose is to "provide a unified system of space data acquisition and delivery to those affected by natural or man-made disasters",[46] the first time the charitable and humanitarian redeployment of the assorted corporate, national space agency, and international satellite assets under its aegis had been used to search for an airliner.[47]

Another 11 countries joined the search efforts by 17 March after more assistance was requested by Malaysia.[48] At the peak of the search effort and before the search was moved to the south Indian Ocean, 26 countries were involved in the search, contributing in aggregate nearly 60 ships and 50 aircraft. In addition to the countries already named, these parties included Australia, Bangladesh, Brunei, Cambodia, China, France, India, Indonesia, Japan, Myanmar, New Zealand, Norway, Philippines, Russia, Singapore, South Korea, Taiwan, Thailand, United Arab Emirates, United Kingdom, United States, and Vietnam.[49][50] While not participating in the search itself, Sri Lanka gave permission for search aircraft to use its airspace.[51] Malaysia deployed military fixed-wing aircraft, helicopters, and ships.[52][53][54] A co-ordination centre at the National Disaster Control Centre (NDCC) in Pulau Meranti, Cyberjaya, was established.[55]

On 16 March, three staff members of the French government agency BEA flew to Kuala Lumpur to share with Malaysian authorities their experience in the organisation of undersea searches, acquired during the search for the wreckage of Air France Flight 447.[56] The United Kingdom provided technical assistance and specialist capabilities from the Ministry of Defence, the UK Hydrographic Office, Department for Transport and the Met Office.[57] The Comprehensive Nuclear-Test-Ban Treaty Organization Preparatory Commission analysed information from its network of infrasound detection stations, but failed to find any sounds made by Flight 370.[58]

On 17 March, Australia agreed to lead the search in the southern locus from Sumatra to the southern Indian Ocean.[81][82] The search would be coordinated by the Australian Maritime Safety Authority (AMSA), with an area of 600,000 km² (230,000 sq mi) between Australia and the Kerguelen Islands lying more than 3,000 kilometres (1,900 mi) southwest of Perth, to be searched by ships and aircraft of Australia, New Zealand, and the United States.[83] This remote area, which Australian Prime Minister Tony Abbott described as "as close to nowhere as it's possible to be", is renowned for its strong winds, inhospitable climate, hostile seas, and deep ocean floors.[84][85] On 18 March, the search of the area began with a single Royal Australian Air Force P-3 Orion aeroplane.[86] On 19 March, the search capacity was ramped up to three aircraft and three merchant ships;[87] the revised search area of 305,000 square kilometres (118,000 sq mi) was about 2,600 kilometres (1,600 mi) south-west of Perth.[88]

Crew members on board a P-8A Poseidon manning terminals while searching for surface debris and locator beacons from Flight 370 in the southern Indian Ocean.

Search efforts intensified on 20 March, after large pieces of possible debris had been photographed in this area four days earlier by a satellite.[89][90][91][92][93] Australia, the United Kingdom, the United States, China, Japan, New Zealand, and South Korea assigned military and civilian ships and aircraft to the search.[94][95] China published images from satellite Gaofen 1 on 22 March that showed large debris about 120 km (75 mi) south west of the previous sighting.[96][97][98] The same day, HMAS Success became the first naval vessel to reach the search area.[99] On 26 March, images from French satellites indicated 122 floating objects in the southern Indian Ocean.[100][101] Thai satellite images published on 27 March showed about 300 floating objects about 200 km (120 mi) from the French satellites' target area.[102] The abundant finds, none confirmed to be from the flight, brought the realisation of the prior lack of surveillance over the area, and the vast amounts of marine debris littering the oceans.[103][104]

On 28 March, revised estimates of the radar track and the aircraft's remaining fuel led to a move of the search 1,100 kilometres (680 mi) north-east of the previous area,[105][106] to a new search area of 319,000 square kilometres (123,000 sq mi), roughly 1,850 kilometres (1,150 mi) west of Perth.[107][108][109][110] This search area had more hospitable weather conditions.[97]

On 30 March, four large orange objects found by search aircraft, described by media as "the so far most promising lead", turned out to be fishing equipment.[111] On 2 April, Royal Navy survey vessel HMS Echo and submarine HMS Tireless arrived in the area,[112] with HMS Echo starting immediately to search for the aircraft's underwater locator beacons (ULBs) fitted to the "black box" flight recorders,[113] whose batteries were expected to expire around 7 April.[114][115] On 4 April, the search area was shifted further north.[116][117]

By 28 April, the drifted search area (outlined in grey) was over 2,500 km (1,600 mi) wide. The dots indicate the probable locations of debris on 28 April, based on a splash point in zones A-B (outlined in blue).

The surface search ended on April 28. In a press conference, Australian Prime Minister Tony Abbott noted that any debris would likely have become waterlogged and sunk and that the aircraft involved in the surface search were "operating at close to the limit of sensible and safe operation".[118] Abbott explained that it was "highly unlikely" that any surface wreckage would be found and, therefore, the aerial searches had been suspended.[118][119][120] The surface search in Southeast Asia and the Indian Ocean lasted 52 days, during 41 of which Australia coordinated the search. Over 4,500,000 km² (1,700,000 sq mi) of ocean surface was searched. In the Southern Indian Ocean, 29 aircraft from seven countries conducted 334 search flights; 14 ships from several countries were also involved.[118][119][120]

On 4 April, the search was refocused to three more northerly areas from 1,060 to 2,100 kilometres (660 to 1,300 mi) west of Learmonth, spanning over 217,000 square kilometres (84,000 sq mi).[116][117] ADV Ocean Shield, fitted with a TPL-25 towed pinger locator, together with HMS Echo—which carried a "similar device", began searching for pings along a 240-kilometre (150 mi) seabed line believed to be the Flight 370 impact area.[114][121][122] Operators considered it a shot in the dark, when comparing the vast search area with the fact that TPL-25 could only search up to 130 square kilometres (50 sq mi) per day.[123]

The Chinese vessel Haixun 01 made a possible ULB detection at 25.975°S 101.461°E / -25.975; 101.461[124] using a handheld hydrophone;[125][126][127][128] the following day, Haixun 01 made another possible ULB detection about 3 km (1.9 mi) further west. HMS Echo and a submarine were later tasked to the location of the Haixun 01's detections, but unable to make any detections.[17]^: 11

Deployment of a towed pinger locator for detecting an aircraft's ULB.

On 6 April, JACC announced that Ocean Shield had also picked up a signal, about 300 nautical miles (560 km; 350 mi) from Haixun 01.[129][130] It was announced the next day that the TPL-25 pinger locator towed by Ocean Shield had picked up a signal twice on 6 April.[131][132] The first was on the morning of 6 April, at approximately 3,000 metres (9,800 ft) depth, and lasted two hours and 20 minutes. The second signal reception took place at approximately 300 metres (980 ft) depth and lasted 13 minutes. During the second episode, two distinct pinger returns were audible. Both episodes of recorded signals, which took place at roughly the same position though several kilometres apart, were considered to be consistent with signals expected from an aircraft's flight recorder ULB.[133] The signals received by Ocean Shield were recorded at the north of a newly calculated impact area, which was announced on 7 April, while the Haixun 01 signals had been recorded at its southern edge.[133][134][135] Ocean Shield detected two more signals on 8 April. The first was acquired at 16:27 AWST and held for 5 minutes, 32 seconds and the second was acquired at 22:17 AWST and held for around seven minutes.[136][137][138] Experts had determined that the earlier signals captured by Ocean Shield were "very stable, distinct, and clear ... at 33.331 kHz and ... consistently pulsed at a 1.106-second interval". These were said to be consistent with the flight recorder ULB.[136] but the frequency of the detections was well outside the manufacturer's specification of 37.5 +/- 1.[139] The later signals were at a frequency of 27 kHz, which raised doubts that they were from a black box.[140] On 10 April, a signal recorded by one of the sonobuoys deployed with a hydrophone at 300 metres depth[141][142] was found unlikely to have originated from Flight 370.[143]

On 14 April, due to the likelihood of the ULBs' acoustic pulses having ceased because their batteries would have run down, the Towed Pinger Locator search gave way to a seabed search using side-scan sonar installed in a Bluefin-21 Autonomous Underwater Vehicle.[144] The first day's search was aborted because the sea bed was considerably deeper than the maximum operating depth of Bluefin. Scanning subsequently resumed[145] and after covering 42 square miles in its first four dives, the submersible was reprogrammed to allow it to dive 604 feet lower than its operational limit of 14,800 feet, when the risk of damage was assessed as "acceptable".[146]

Ocean Shield deploys Bluefin-21, 14 April 2014.

Bluefin-21 required 16 missions to complete its search of the 314 square kilometre area around the detections made by the Towed Pinger Locator.[147][148] The seafloor sonar search was suspended on 2 May for the AVD Ocean Shield to return to port to replenish supplies and personnel.[149] Within two hours of its first deployment after returning to the search area on 13 May, Bluefin-21 developed a communications problem and was recovered.[150] Spare parts from the UK were required and the AVD Ocean Shield returned to port to collect the parts.[151][152] After Bluefin-21 was repaired, the seafloor sonar survey resumed on 22 May.[153][154]

The seafloor sonar survey was completed on 28 May. After 30 deployments of the Bluefin-21 to depths of 3,000–5,000 m (9,800–16,400 ft), which scanned 860 km² (330 sq mi) of seabed, no objects associated with Flight 370 were found. The following day, after analysis of data from the last mission, the ATSB announced the search in the vicinity of the acoustic detections was complete and the area could be discounted as the final location of Flight 370.[155][156] The announcement followed a statement by U.S. Navy's Deputy Director of Ocean Engineering that all four pings were no longer believed to have come from the aircraft's flight recorders.[157] Truss informed parliament that, beginning in August, after a new commercial operator for the search effort had been selected, the search would move into a new phase "that could take twelve months".[158] Equipment envisaged to be used would include towed side-scan sonar.[159]

Examination of possible end-of-flight scenarios indicate the aircraft may be located relatively close to the seventh BTO arc. On account of this, the underwater search began at the seventh BTO arc and has progressed outwards.[175]^: 12 Towed underwater vehicles equipped with synthetic aperture sonar, side-scan sonar, and multi-beam echo sounders are towed close to the seafloor to create a three-dimensional picture of the seafloor topography.[191]

The sonar data was analysed on board the search vessels at the time of acquisition by sonar analysts and a representative of the ATSB. Data from Fugro vessels was transmitted by satellite to a team of Fugro sonar specialists and geophysicists in Perth then transmitted to the ATSB. Data from the ProSAS sonar used by the GO Phoenix and Dong Hai Jiu 101 was analysed onboard and transmitted to the ATSB during port calls in Perth. In both cases, data was further reviewed by ATSB staff in Canberra and both a quality assurance review and independent expert review were performed by experts in the United States. At all stages, the data was analysed for quality, coverage, and contacts (areas of interest).[192]^: 11 Contacts were classified into three levels:[6]

• Category 3: Areas that stand out from their surroundings and are of some interest, but are unlikely to be significant to the search.
• Category 2: Areas of comparatively more interest, but still unlikely to be significant to the search. Some of these are possibly manmade objects, such as objects with the dimensions of shipping containers.
• Category 1: Areas of high interest to the search, requiring prompt investigation.

Through 20 December 2016, 605 Category 3, 39 Category 2, and two Category 1 contacts were identified. The two Category 1 contacts were identified as a shipwreck and a rock field.[192]^: 11

Fugro Discovery departing from Fremantle, February 2016

• MV Fugro Discovery initially carried an EdgeTech DT-1 towfish, named Dragon Prince,[191] which was replaced with a new towfish, named Intrepid, in April 2015.[193] The vessel joined the underwater search on 23 October 2014.[166]
• MV Fugro Equator was scheduled to join the search in late October 2014,[166] but resumed bathymetric survey work from 19 November to 17 December while waiting for equipment to arrive.[194][195] It returned to Fremantle for mobilisation for the underwater search and joined the underwater search on 15 January 2015.[180] The Fugro Equator was the last vessel involved in the underwater search when it departed the search area on 17 January 2017—the day the search was officially suspended.[189][196]
• MV Fugro Supporter participated in the search from late January until early May 2015.[179][181][186]
• MV GO Phoenix carried experts and equipment from Phoenix International.[166] The first vessel to begin underwater search operations, on 6 October 2014.[176] It left the search on 20 June 2015.[187]
• MV Havila Harmony carried an AUV used to search the most rugged terrain. It arrived in the search area on 5 December 2015.[197][198]
• MTS Dong Hai Jiu 101 was a rescue and salvage tug supplied by the Chinese government. The vessel joined the search in February 2016; it left the search area on 3 December 2016.[199] In September 2016, Australian media pointed out that the Dong Hai Jiu 101 spent few days participating in the search—an estimated 17–30 days between February and September—but spent a significant amount of time anchored in or near the port of Fremantle; the media suggested that the Chinese vessel was likely engaged in espionage—signals intelligence gathering and monitoring of Australian and allied naval traffic—around the port of Fremantle.[200][201] When active in the search, it carried a remotely operated underwater vehicle (ROV) from Fremantle to recover the towfish lost in March and was mobilised from October through December with an ROV to participate in the search.[199][202][203]

MV Fugro Discovery, MV Fugro Equator, and MV GO Phoenix were equipped with towed underwater vehicles (also known as "towfish"), which carry synthetic aperture sonar, side scan sonar, and multi-beam echo sounders. The towfish were towed behind the vessel on cables up to 10 kilometres (6.2 mi) in length. The length of the cable was varied to keep the towfish at an altitude of 100–150 metres (330–490 ft) above the seafloor. The data collected by the instruments was relayed in real-time along the tow cable to the vessel, where it was processed and analysed to see if any debris associated with Flight 370 was present on the seafloor.[191]

The MV Fugro Supporter and MV Havila Harmony were equipped with a Kongsberg HUGIN 4500 autonomous underwater vehicle (AUV), which carried the same instruments as the towfish deployed by the other vessels. Unlike the towfish, the AUVs were self-propelled and could manoeuvre themselves to maintain a constant altitude above the seafloor. When launched, the torpedo-shaped AUV dived to the proper altitude, then followed a pre-programmed search pattern for up to 24 hours, collecting sonar data for up to 24 hours; when the mission was complete, the AUV surfaced and was recovered by the vessel, where the data was downloaded and the batteries were swapped for the next mission. The AUVs were used to scan areas which could not be effectively searched by the towfish on the other three vessels.[191][198] Because of the difficulty in operating the AUV during the austral winter months due to sea conditions, the AUV operations were suspended during the austral winter months.[204]

In January 2016, a towfish pulled by the Fugro Discovery collided with an underwater mud volcano despite its use of forward-looking obstacle avoidance sonar.[205][206] The towfish was later recovered by the Havila Harmony and transferred to Fugro Discovery to continue the search.[207] Another towfish was lost on 21 March 2016.[208]

On 13 May 2015, it was announced that Fugro Equator had discovered wreckage of a previously uncharted 19th century cargo ship 4,000 metres underwater, more than 1,000 km off Australia's west coast.[209][210] Fugro Supporter was diverted to the area to deploy an unmanned submarine to scan the seabed. Imaging of the site clearly showed an anchor and other manmade objects. It was announced that the related imagery would be provided to expert marine archaeologists for potential identification.[211] Peter Foley, director of search operations, called it a "fascinating find but it's not what we're looking for".[212]

In December 2015, an anomalous sonar contact was made of a possible man-made object and the Havila Harmony was sent to investigate. On 2 January 2016, its AUV captured high-resolution sonar images, which confirmed the contact was a shipwreck. Experts at the Western Australian Museum advised that the ship is likely an iron or steel vessel from the turn of the 19th century.[213]

A week after the discovery of a flaperon from Flight 370 on a beach on Réunion, France announced plans for an aerial search for possible marine debris around the island. On 7 August 2015, France began searching an area 120 km (75 mi) by 40 km (25 mi) along the east coast of Reunion.[225] Foot patrols for debris along beaches are also planned.[226] Malaysia asked authorities in neighboring states to be on alert for marine debris which could be from an aircraft.[227]

On 23 March 2016, what appeared to be a piece of a Rolls-Royce engine cowling was found on a beach in South Africa. The Malaysian authorities sent a team to recover the piece and determine whether or not it came from the missing plane.[228]

On 24 June 2016, the Australian Transport Minister, Darren Chester, said that a piece of aircraft debris was found on Pemba Island, off the coast of Tanzania. The Australian Transport Safety Bureau announced on 22 June that a piece of debris found on an Australian island earlier in June was not from Flight 370.[229] On 20 July, the Australian government released photos of the piece of debris found on Pemba Island around 24 June, believed to be an outboard wing flap.[230]

Surface Velocity Program drift buoy, as used by the GRP

The proposed additional searching was to be based on the drift modelling work undertaken by the CSIRO in 2016, and presented in the CSIRO Drift Reports 2, 3 and 4 subsequently published by ATSB.[236] The essence of the CSIRO work was to investigate the most likely paths taken by drifting debris originating along the 7th BTO arc search area, and comparing this to recovered debris along the east coast of Africa and neighbouring Indian Ocean islands. Of particular interest was the first main recovered object, the starboard flaperon, which was found on Reunion Island in July 2015.[237]

Early drift analysis, based on data collected over years from hundreds of buoys deployed by the Global Drifter Program of the US National Oceanic and Atmospheric Administration, worked on the presumption that currents in the southern portion of the Indian Ocean, around 40-50 degrees South latitude, flow to the east accompanied by predominately westerly winds. Currents and winds in the northern portion however, around 20 degrees South, both flow to the west. Between these zones lies an "intermediate" area with no dominant movement either way. The main MH370 underwater search areas, from around 36°S to 40°S, extended across several of these zones, making debris drift prediction uncertain.[237]

Satellite oceanographic data obtained by the CSIRO team revealed that in early March 2014, a ridge of high sea level cut across the main 7th arc search area, producing two prominent westward surface currents in the "intermediate" zone, one at 30°S and one at 35°S. This would have initiated a westerly movement of any floating debris in these areas. This helped address one of the key questions raised by the analysts as to why no debris was being found along the nearby west coast of Australia. This absence, along with the satellite data, strengthened the case for the impact position being further located north than previously calculated, with a latitude around 35.6°S 92.8°E / -35.6; 92.8 (CSIRO crash area) being the most likely according to the data.[237]

High-resolution satellite images from a number of sources (Pleiades,[238] NOAA SST[239]) indicated a number of possibly man-made floating objects which, when analysed in conjunction with the revised drift analysis, could potentially have all originated in close proximity within the 7th BTO arc. The conclusion of the CSIRO Drift Report 3 was that this location was most likely to be north of the original search area, at around 35°S latitude.[240]

Boeing 777 flaperon

The recovery of the Boeing 777 flaperon at Reunion Island in July 2015 was unexpected by the prevailing drift analyses, as it appeared to arrive far earlier than was predicted for a crash location within the search area.[241] Physical testing of both replica and real flaperon items was undertaken to try to explain this occurrence.[237]

A genuine 777 flaperon was obtained from the ATSB by the CSIRO and cut down to resemble the recovered item. Buoyancy checks were done to correctly match that of the MH370 debris, and two weeks of fields tests were undertaken in North West Bay and Storm Bay, Tasmania to measure the flaperon's wind and current drift under a range of conditions. This testing revealed several significant pieces of information. Firstly, the flaperon, when cut to the shape of the recovered item, didn't drift exactly in-line with the prevailing wind, but exhibited a leeway drift of around 20° to the left. Secondly, due to the high float attitude of the flaperon, the wind effect was greater than originally expected, resulting in a corresponding overall faster speed of drift. These two data provided concurrence between the July 2015 arrival of the flaperon on Reunion Island, and it having had a source location in the northern end of the original search area.[237]

MH370 7th arc search area, with focus of 2018 search noted

Simulated drift trajectories of a large number of such flaperons were mapped in the analysis, based on them all having a source location within the southern end of the "remaining high probability area" identified in the ATSB First Principles Review (FPR) report.[242] This area extended approximately 25 nautical miles (46 km; 29 mi) each side along the 7th arc from around 32.5°S in the north to around 36°S in the south. The modelled trajectories indicated that a source location of around 35°S was consistent both with the maximum probability that no debris would be found on Western Australia, and good consistency that debris would appear at Reunion Island from July 2015.[237]

The reports concluded that the most likely location of MH370 is within the southern end of the 25,000 square kilometres (9,700 sq mi) area identified in the ATSB FPR, near 35°S.[241]

Deep-water autonomous underwater vehicle

In January 2018, Ocean Infinity announced its intention to resume the search in the narrowed 25,000 km² area. The attempt was approved by the Malaysian government, provided payment will be made only if the wreckage is found.[243][244] Initially, the search was to be focused within the "high-probability" area identified by the ATSB.[245] Under the contract, Ocean Infinity was to be paid up to $70 million, depending on the duration of the search; however, the plane had to be found in 90 days after the start of the search.[246] In March 2018, the Malaysian government clarified that the 90 days limit did not include the necessary breaks for the vessel to resupply, and so the search was to continue until mid-June.[247]

Ocean Infinity chartered the Norwegian multi-purpose offshore vessel Seabed Constructor to conduct the search. Maritime assets deployed by Ocean Infinity for the search include up to eight autonomous underwater vehicles (AUVs), each able to operate independently within a group, at depths down to 6,000 metres (20,000 ft) without physical tethering to a surface vessel. The AUVs were to be equipped with a range of scanning technologies including side-scan sonar, multi-beam echo sounder, bottom profiler, high-definition camera, environmental sensors, magnetometer and synthetic aperature sonar.[245][248][249]

The vessel arrived at the search area on 21 January 2018 and started the search on 22 January 2018. The planned search area of site 1, where the search begun, was 33,012 square kilometres (12,746 sq mi), while the extended area covered some more 48,500 square kilometres (18,700 sq mi).[250] During the search, operations were temporarily suspended three times, while Seabed Constructor transited to Fremantle for refuelling, resupply, and crew change (twice); each transition took some ten days.[251][252][253][254][255][256] In March 2018, Seabed Constructor started to search small areas beyond the originally planned search area.[257] In April 2018, all areas of site 1 has been searched. The search continued northeast,[257][258] and later that month area further northeast along the 7th arc (Site 4) was added to the search plan.[259] In May 2018, the total area covered during the search exceeded 81,500 square kilometres (31,500 sq mi), roughly the originally planned area of site 1 together with the extended area.[256] The total area searched was some 120,000 square kilometres (46,000 sq mi).[260]

According to Loke Siew Fook, the Malaysia's new transport minister, the search was to end 29 May 2018.[261] As of 1 June 2018, the search was reported as ongoing, searching beyond site 4 in the area where in 2014 MV Haixun 01, operated by the China Maritime Safety Administration, had detected an acoustic signal[262] suspected to originate from an underwater location beacon attached to the plane's black boxes. However, while several detections had been made, none could have been detected when the ship had passed the same location on an opposing heading, and independent analyses of the detections had determined that the signals had not match the performance standards of the underwater location beacons. It had been considered possible but unlikely that the signals could have had originated from a damaged one.[17]^: 13 The spokesperson for Ocean Infinity confirmed the contract with the Malaysian government had ended and the search itself would end soon,[263] probably on 8 June 2018.[264]

On 9 June 2018, it was reported that the search had ended unsuccessfully.[265] Ocean floor mapping data collected during the search were donated to the Nippon Foundation GEBCO Seabed 2030 Project to be incorporated into the global map of the ocean floor.[266][260]

In March 2019, in the wake of the 5th anniversary of the disappearance, the Malaysian government stated it was willing to look at any "credible leads or specific proposals" regarding a new search.[267][268] Ocean Infinity stated they are ready to resume the search on the same no-fee no-find basis, believing they would benefit from the experience they gained from their search for the wreck of Argentinian submarine ARA San Juan and bulk carrier ship Stellar Daisy, with the most probable location still being somewhere along the 7th arc around the area identified previously, the one on which their 2018 search was based.[269] In March 2022, Ocean Infinity committed to resuming its search in 2023 or 2024, pending approval by the Malaysian government, with two new robotic ships to replace Seabed Constructor.[13]

On 8 March 2023, the ninth anniversary of the disappearance of flight MH370, Malaysian Transport Minister Anthony Loke vowed not to "close the book" on MH370, adding that due consideration would be given to future searches if there was "new and credible information" on the aircraft's potential location.[270]

If Flight 370 had impacted the ocean hard, resulting underwater sounds could have been detected by hydrophones, given favourable circumstances.[17]^: 40[271][272] Sound waves can travel long distances in the ocean, but sounds that travel best are those that are reflected into the 'deep sound channel' usually found between 600–1,200 m beneath the surface. Most of the sound generated by an aircraft impacting the ocean would travel straight down to the seabed, making it unlikely that any of these sounds would be reflected into the deep sound channel unless the seabed sloped. Sounds from pieces of the aircraft imploding at depth would be more likely to travel in the deep sound channel. "The combination of circumstances necessary to allow [detection of an ocean impact] would have to be very particular", according to Mark Prior, a seismic-acoustic specialist at the Comprehensive Nuclear-Test-Ban Treaty Organization (CTBTO), who also explains that "given the continuing uncertainty regarding the fate of MH370, underwater acoustic data still has the possibility of adding something to the search."[272] When an Airbus A330 hit the Atlantic Ocean at speed of 152 kn (282 km/h; 175 mph), no data relating to the impact was detected in hydroacoustic recordings, even when analysed after the location of that aircraft was known.[272][273] As with the analysis of the Inmarsat satellite data, the hydroacoustic analysis uses the data in a way very different from that originally intended.[273]

The Australian Transport Safety Bureau requested the Curtin University Centre for Marine Science and Technology (CMST) analyse these signals.[17]^: 40 Scientists from CTBTO and Geoscience Australia have also been involved with the analysis. Available sources of hydroacoustic data were:[17]^: 40, 47[271][272][273]

• The Comprehensive Nuclear-Test-Ban Treaty Organization (CTBTO), which operates a system of sensors to detect nuclear tests to ensure compliance with the Comprehensive Nuclear Test Ban Treaty. Data was analysed from CTBTO hydrophones located south-west of Cape Leeuwin, Western Australia (HA01) and in the northern Indian Ocean. These stations have two hydrophones each, separated by several kilometres, allowing a bearing to be calculated for the source of noise to within 0.5°.
• Australia's Integrated Marine Observing System (IMOS).[274] Data from an acoustic observatory (RCS) 40 km west of Rottnest Island, Western Australia, near the Perth Canyon. IMOS stations have just one hydrophone each and therefore cannot provide a bearing on the source of the noise. Several IMOS recorders deployed in the Indian Ocean off northwestern Australia by CMST may have recorded data related to Flight 370. As of 2014, these recorders were not recovered as part of the investigation. These sensors record only five minutes out of every fifteen and are likely to be contaminated by noise from seismic surveys.
• It is unclear what other sources of hydroacoustic data are available in the region. India and Pakistan operate submarine fleets, but the JACC claims they are not aware of any hydrophones operated by those countries. The U.S. Navy operated a vast array of hydrophones—the Sound Surveillance System (SOSUS)—during the Cold War to track submarines, which is believed to remain in operation. Asked if any SOSUS sensors are located in the Indian Ocean, a spokesman for the U.S. Navy declined to comment on the subject, noting that such information is classified.

Scientists from the CTBTO analysed their recordings soon after Flight 370 disappeared, finding nothing of interest. However, after the search for the flight shifted to the Indian Ocean, CMST collected recordings from the IMOS and found a clear acoustic signature just after 01:30 UTC on 8 March.[271] This signature was also found, but difficult to discern from background noise, in the CTBTO recordings from HA01, likely because HA01 receives a lot of noise from the Southern Ocean and Antarctic coastline.[271]

Map of the Indian Ocean. Dots off the southwest corner of Australia indicate the location of hydrophone stations. The seventh arc and region of the underwater search is depicted west of Australia. The area of uncertainty where the hydroacoustic detection originated is a long, slim region in red running from the center of the coast of Arabia on the Indian Ocean to a point in the middle of the Indian Ocean, but far from the seventh arc and underwater search area.

The CMST researchers believe that the most likely explanation of the hydroacoustic data is that they come from the same event, but unrelated to Flight 370.[17]^: 47 They note that "the characteristics of the [event's acoustic signals] are not unusual, it is only their arrival time and to some extent the direction from which they came that make them of interest".[17]^: 47 If the data relates to the same event, related to Flight 370, but the arc derived from analysis of the aircraft's satellite transmission is incorrect, then the most likely place to look for the aircraft would be along a line from HA01 at a bearing of 301.6° until that line reaches the Chagos-Laccadive Ridge (approximately 2.3°S, 73.7°E). In the latter possibility, if the acoustic recordings are from the impact of the aircraft with the ocean, they likely came from a location where water is less than 2,000 m deep and the seabed slopes downwards towards the east or southeast; if they came from debris imploding at depth, the source location along this line is much less certain.[17]^: 47 The lead CMST researcher believed the chance the acoustic event was related to Flight 370 to be very slim, perhaps as low as 10%.[275] The audio recording of the suspect detection was publicly released on 4 June 2014;[271][273] the ATSB had first referenced these signals in a document posted on its website on 26 May.[273]

The aircraft's flight recorders were fitted with Dukane DK100 underwater acoustic beacons—also known as "underwater locator beacons" (ULBs) or "pingers"—which are activated by immersion in salt water and thereafter emit a 10 millisecond pulse every second at a frequency of 37.5±1 kHz. The beacons are limited by battery life, providing a minimum of 30 days and have an estimated maximum life of 40 days, according to their manufacturer. The nominal distance at which these beacons can be detected is 2,000–3,000 metres.[17]^: 11 Because the flight recorders to which they are attached could provide valuable information in the investigation, an intense effort was made to detect the beacons' pings before their batteries expired.

HMS Echo made one possible detection on 2 April—the same day the ship joined the search effort. The following day, following tests, the detection was dismissed as an artefact of the ship's sonar system.[17]^: 11[113] On the afternoon of 5 April Perth time, HMS Echo detected a signal lasting approximately 90 seconds. The second detection was made within 2 km from the first detection.[276]

MV Haixun 01, operated by the China Maritime Safety Administration, detected a signal at 37.5 kHz pulsing once per second on 4 April and again on 5 April at a position 3 km west of the first detection.[130] HMS Echo was sent to the location of the MV Haixun 01 detections and determined that the detections were unlikely to originate from ULBs attached to the plane's black boxes due to the depth of the seafloor, surface noise, and the equipment used. A submarine sent to the location made no acoustic detections.[17]^: 13

The towed pinger locator on the deck of a vessel. An inset shows a rack of electrical equipment.

ADV Ocean Shield was sent to the search area with two Phoenix International TPL-25 towed pinger locators (also known as "towfish"). Shortly after one of the towfish was deployed, while descending, an acoustic signal was detected at a frequency of 33 kHz on 5 April. Further detections were made on 5 April and on 8 April, but none could be detected when the ship passed the same location on an opposing heading.[17]^: 12

Independent analyses of the detections made by ADV Ocean Shield determined that the signals did not match the performance standards of the ULBs attached to the aircraft's black boxes. However, although unlikely, they noted that the signals could have originated from a damaged ULB.[17]^: 13

Between 6–16 April, AP-3C Orion aircraft of the Royal Australian Air Force deployed sonobuoys, which sank to a depth of 300m to detect the acoustic signature of the ULBs attached to the aircraft's black boxes. Sonobuoy drops were carried out at locations along the calculated arc of the final satellite communication with Flight 370 where seafloor depths were considered favourable, near the MV Haixun 01 detections, and along the bearing determined by the Curtin University research team of a possible impact event. One AP-3C Orion sortie was capable of searching an area of 3,000 square kilometres (1,200 sq mi). No acoustic detections related to the ULBs attached to the aircraft's black boxes were made by the sonobuoys.[17]^: 13

The interim report released by the Malaysian Ministry of Transport in March 2015 mentioned, for the first time publicly, that the battery for the ULB attached to the flight data recorder had expired in December 2012. The battery's (and thus the ULB's) performance may have been compromised, but this likely was not significant in the search, given the close range at which the detection must be made and the vast search area.[277][278]

• MH 370 Preliminary Report – Preliminary report issued by the Malaysia Ministry of Transport. Dated 9 April 2014 and released publicly on 1 May 2014.
• Factual Information: Safety Investigation for MH370 Archived 9 March 2015 at the Wayback Machine – Interim report released 8 March 2015 (584 pages).
• MH370 – Definition of Underwater Search Areas (2014) – Report by the Australian Transport Safety Bureau, released 26 June 2014, and the most comprehensive report on Flight 370 publicly released at that time. The report focuses on defining the search area for the fifth phase, but in doing so provides a comprehensive overview/examination of satellite data, the failed searches, and possible "end-of-flight scenarios".
• MH370 – Definition of Underwater Search Areas Archived 12 April 2021 at the Wayback Machine (2015) – Report by the Australian Transport Safety Bureau, released 3 December 2015, covering the Bayesian method analysis made by Australia's Defence Science and Technology Group and other developments since mid-2014 in defining the search area.
• MH370 – Search and debris examination update Archived 8 October 2022 at the Wayback Machine (2016) – Report by the Australian Transport Safety Bureau, released 2 November 2016, comprising further analysis of satellite data, additional End of Flight simulations, analysis of flight debris (wing flap), and enhanced debris drift modelling.
• MH370 – First Principles Review Archived 30 December 2016 at the Wayback Machine – Report by the Australian Transport Safety Bureau, released 20 December 2016, documenting the proceedings and outcomes of the First Principles Review meeting held in Canberra between 2–4 November 2016. The review identified a previously unsearched area of 25,000 square kilometres (9,700 sq mi) as having the highest probability of containing the aircraft wreckage.
• Search for MH370 and ocean surface drift – part III – Report to the Australian Transport Safety Bureau by the CSIRO, of 26 June 2017 and released 16 August 2017, comprising analysis of multiple surface objects identified by the Geosciences Australia report of 19 July 2017, as to whether such objects could have originated within the established MH370 search areas
• Minchin, S.; Mueller, N.; Lewis, A.; Tran, M. (16 August 2017). "Summary of imagery analyses for non-natural objects in support of the search for Flight MH370" (PDF). www.atsb.gov.au. Geoscience Australia. Archived from the original (PDF) on 18 August 2017. Retrieved 18 August 2017.
• Summary of imagery analyses for non-natural objects Archived 18 August 2017 at the Wayback Machine – Report by Geoscience Australia, released 16 August 2017, comprising analysis of imagery from data from the PLEIADES 1A satellite, of floating objects identified in the southern Indian Ocean.
• The Operational Search for MH370 (Final) Archived 10 March 2020 at the Wayback Machine – The final report from the Australian Transport Safety Bureau (ATSB), released 3 October 2017, and documenting where and how the search for MH370 was conducted, the results obtained and analysis for where future underwater searches could be undertaken. It concludes that the reasons for the loss of MH370 cannot be established with certainty until the aircraft is located.

• Australian Maritime Safety Authority
• Australian Transport Safety Bureau Archived 8 October 2014 at the Wayback Machine
• Malaysia Airlines Archived 22 May 2020 at the Wayback Machine
• Malaysian Ministry of Transport Archived 11 November 2020 at the Wayback Machine
• US Department of Defense