Hours after Malaysian Air Flight 370 vanished on March 8, Inmarsat Plc assembled a team of engineers thousands of miles away at London headquarters for a marathon data-crunching session to help find the missing jet.
They had little more to work with than a half-dozen vague signals to a company satellite and a lot of assumptions. By March 11, they calculated two plausible trajectories far from the intended path to Beijing from Kuala Lumpur -- one north into central Asia and one south toward Antarctica. While the latter was more probable, they couldn’t be sure.
“At that point, they felt we’d gone as far as we could,” said Chris McLaughlin, a spokesman for Inmarsat.
Yet knowing how implausible it would be to search two enormous zones, and that there was little chance of new evidence emerging, the scientists and engineers at Inmarsat vowed to try again. Their breakthrough, while only narrowing the possibilities, let them rule out the northern path and prompted Malaysia’s prime minister to declare the Boeing Co. 777-200ER jet lost in the Indian Ocean with all 239 passengers and crew presumed dead.
With the plane’s communications systems having been shut off and no wreckage to be found, the engineers’ conclusions have been the closest thing to a resolution of the longest disappearance in modern airline history.
“They worked together for six or seven days straight,” said McLaughlin, with the international team unwinding at the in-house gym during breaks or fetching pizza to sustain the round-the-clock mission.
“What we discovered was that the northern projected path had no correlating pings appear on it, while for the southern projected path, we had frankly an absolute correlation. The plotted positions, the plotted lines all matched each other.”
The group narrowed its initial calculations using principles of the so-called Doppler effect, named after 19th-century Austrian physicist Christian Doppler, who explored how movement can alter a signal profile.
The idea was to look at the velocity of the aircraft relative to the satellite. Depending on this relative movement, the frequency received and transmitted differs from its normal value in the same way that the sound of a passing train changes as it approaches and then moves away.
About half a dozen engineers and scientists analyzed the difference between the frequency the ground station expected to receive and what was actually measured. They then compared the analysis with the flights of other Boeing 777s that had traveled on both the north and south arcs to see which patterns matched up.
While on the ground at Kuala Lumpur airport, and during the early stage of the flight, MH370 transmitted several messages, Malaysian acting Transport Minister Hishammuddin Hussein said yesterday. Data received during the flight help determine what direction the plane may have flown, he said.
Based on the aircraft speed, it was then possible to estimate positions when the last “complete handshake,” or communication, took place. There is evidence of a partial handshake between the aircraft and the ground station eight minutes after the complete communication, Hishammuddin said.
“The new analysis was convincing enough for the AAIB to brief the prime minister that the aircraft flew in the southern corridor,” he told a press conference yesterday. “This type of analysis has never been done in an investigation.”
The Inmarsat teams, meeting in the glass-coated company headquarters that borders London’s financial district, also used Boeing data on aircraft performance to see where the Doppler effect would result in a pattern that matched the data from Flight 370, McLaughlin said. Boeing declined to comment.
The team’s “eureka moment” came over the weekend, McLaughlin said. To validate its approach, the team shared its findings with engineers familiar with satellites as a form of scientific peer review. He declined to identify the company.
The review group concurred on March 23. Inmarsat took its findings “with some degree of nervousness” to the U.K. Air Accidents Investigation Branch, which presented them to Malaysian authorities heading the search, McLaughlin said.
Inmarsat was set up in 1979 as an intergovernmental organization to provide satellite communications for ships. It operates three constellations of 10 satellites in geostationary orbit 35,786 kilometers (22,200 miles) above the Earth.
The company counted Deutsche Telekom AG, Lockheed Martin Corp. and BT Group Plc among early investors before private-equity firms Apax Partners LLP and Permira Advisers LLP Ltd. took it over in 2003.
Inmarsat sold shares for the first time in 2005, and is valued at about 3.26 billion pounds ($5.38 billion). Other satellite companies competing in air-traffic services include Globalstar Inc., based in Covington, Louisiana, and Iridium Communications Inc., whose headquarters are in McLean, Virginia.
The plane probably ran out of fuel west of Perth, Australia, Malaysian Airline System Bhd. Chief Executive Officer Ahmad Jauhari Yahya told reporters yesterday at the press conference.
The AAIB declined to comment other than to confirm it worked with Inmarsat.
The mathematical modeling has played a crucial role in narrowing the scope of a search mission now in its third week. The search along southern Indian Ocean resumed today with ships and aircraft from six nations scouring the waters.
While countries from Australia to China have rushed planes and ships to an area about 2,500 kilometers southwest of Perth, bad weather, high seas and the logistical challenges of reaching one of the most remote spots on Earth have hampered the efforts.
Authorities have said that Inmarsat would keep working to whittle down the search area.
“That’s the best they can do with the information available to them,” said Robert Mann, a former American Airlines executive and aviation consultant based in Port Washington, New York. “That’s the limitation, as you’re forced to innovate.”