For a while, it seemed as if the search for missing Malaysia Airlines Flight 370 was unfolding the way that these things generally do: a few tantalizing clues were about to yield tangible evidence, the wreckage would be located, the black box data would be analyzed, and the mystery would be resolved. On March 15, the Malaysian prime minister announced that analysis of pings transmitted from the plane to an Inmarsat satellite showed that it had stayed in the air for eight hours and must have wound up on one of two arcs, one which ran through Central Asia and the other which cut across the Indian Ocean. He later clarified that further analysis had proven that the plane must be in the southern arc, and that there was therefore no hope that anyone aboard had survived. On March 20, Australian Prime Minister Tony Abbott assured his nation’s parliament that satellite imagery of debris southwest of Perth presented “new and credible information” about the fate of the plane. By working backward from the wreckage’s location using known drift patterns, searchers would be able to lay out a grid on the seabed in which to trawl with listening devices for the acoustic pinger.
Then progress stalled. Searchers were never able to locate the debris that looked so promising in satellite imagery. Nor were they able to locate other hot leads beamed down from space. Search aircraft and crew, who were battling bad weather, didn’t find anything that looked like it came from a plane. And then, on March 28, the Australian Maritime Safety Authority announced that it was abandoning the original search area in favor of another one farther to the north. Why? As I’ve explained elsewhere, investigators had settled on that first search area after a making a baseline assumption that the plane was traveling at a steady speed of about 450 knots, on a more-or-less straight heading; when that didn’t pan out, they decided to try a search area corresponding to a slower flight speed. In other words, the much-vaunted “credible information” had only ever been a hopeful guess.
In announcing the new search area a few days ago, the authorities painted the change as a move forward. Malaysia’s transport minister called the new search area “more focused than before.” But in fact it was much bigger, about the size of New Mexico. Instead of converging on a solution, searchers were finding themselves facing ever-bigger challenges. Each day that went by meant that whatever wreckage was still floating was becoming more diffuse, and hence less useful in figuring out where the plane went down. And time was running out: Soon the batteries in the plane’s acoustic pinger will reach the end of their mandated lifespan.
In the absence of physical evidence that MH370 is in the southern Indian Ocean, many are focusing their attention on the one actual clue that we have about the plane’s fate: those Inmarsat pings. On the Internet, the technically savvy have banded together to dissect what little information Inmarsat has released about the pings and its analysis of them. By reversing-engineering this data, they hope both to reconstruct the logic used to obtain the current search areas, and ultimately to crowdsource a more comprehensive analysis that could provide a definitive answer for where the plane could and couldn’t have gone. Among the movement’s leading lights is space scientist Duncan Steel, a New Zealand resident who has posted elaborate analyses of satellite motion and Doppler shift on his blog. “As a hard-bitten scientist,” he told me via email, “I try to cut everything back to the few things we actually know, and then methodically work forwards from there.”
One of the things that’s become clear, unfortunately, is just how difficult it is to draw any firm conclusions from the ping data. Even the confident assertion by the Malaysian prime minister that the northern route had been ruled out seems, on closer inspection, to be a fairly rickety claim. What Inmarsat actually said in its analysis was that if one assumed that the plane maintained a steady speed, then a straight-ahead course on the southern arc more nearly matched the ping data than a similar course to the north. But we already know that before it disappeared the plane was changing speeds and flying a zigzag course. (Indeed, to get to the southern Indian Ocean, MH370 would have had to do an about-face from its last known course; if it had followed a straight-ahead track, it would have wound up in Kazakhstan.) The “evidence” that the Malaysian prime minister cited as justification for telling passengers’ families that all hope has been lost was really just the output of a model into which an assumption had been fed. Or perhaps there’s more to it than that—we just don’t know, because neither Inmarsat nor the Malaysian authorities will share their full data or show their work.
As days go by with no wreckage turning up, the authorities are maintaining their resolute stance; Abbott said on March 31 that the tempo of search operations "is increasing, not decreasing.” But time is not on our side. Can we afford to be so confident in our assumptions? I’m not saying that I believe that MH370 is in Central Asia (though I have previously outlined the case that could be made for that view, and was much criticized for it). I am saying that the evidence that MH370 is in the Indian Ocean is very shaky indeed. I’m criticizing not the idea that it’s in the Indian Ocean, but the confident assertion that it must be.
Until someone fishes a seat cushion or luggage tag out of the sea, the only way we’re likely to understand more about the fate of MH370 is if Inmarsat and the Malaysian authorities open up the ping data to the public for independent analysis. So far, they’ve been reluctant to lift their veil of secrecy. But if the public—and especially the passengers’ families, who have been told to accept the death of their loved ones on the basis of that secret data alone—demand its release with sufficient force, perhaps they will change their tune.