When the black boxes of Air France Flight 447 were finally located two years after the plane crashed in the Atlantic Ocean halfway between Rio de Janeiro and Paris, investigators found them in an area that was carefully searched early on and then abandoned.
"Multiple countries searched that area," says Richard Limeburner, a physical oceanographer at Woods Hole Oceanographic Institution in Woods Hole, Massachusetts. "I'm not sure there was someone in charge at that point. They also had bad weather, with some windy conditions, so they couldn't see very far."
Limeburner and another oceanographer eventually found the wreckage, guided by satellite data collected from drift buoys fitted with sensors and used to track ocean currents.
Oceanographers in Australia hope the same technique will lead them to Malaysia Airlines Flight 370 which disappeared March 8 and is believed to have crashed in the Indian Ocean west of Australia.
Since March 18, a team of oceanographers at CSIRO, Australia's national scientific research agency, has been analyzing computer models that predict possible crash zones. Using satellite images of floating debris spotted in those zones, the team then calculates the debris drift in order to direct ships and planes to it. (See "How Finding Debris Could Quickly Solve Mystery of Malaysian Airlines Flight.")
If debris from the plane is found, the Australians would next track it back to the point it entered the ocean, using a method known as "hindcasting." If the calculations are correct—or at least close—the largest, heaviest pieces of wreckage would lie below that point, or nearby.
So far, none of the debris retrieved in three separate zones belongs to the missing Malaysian plane, according to the Australian Maritime Safety Authority, which has been overseeing the search.
Analyzing the Variables
"Those are fairly large impact zones," says David Griffin, one of two CSIRO researchers heading a team of ten scientists and engineers combing through the data. "So once we decide, okay, if something landed in Zone 3 on March 8, where could it possibly be now? It comes down to: How fast does material disperse in the ocean?"
That depends on many variables. Griffin, a physical oceanographer, and his team must consider weather, waves, random turbulence, and the buoyancy of the objects themselves, as well as ocean and wind currents, which run in opposite directions. Heavier objects that are partially submerged, for example, drift with ocean currents and could travel in the opposite direction from lighter objects, such as Styrofoam cups or life preservers, which would bounce along the surface and be carried by the wind.
"It's very, very difficult to track things in the ocean," says Limeburner, who used hindcasting and drifting buoys to find a steamship off Nantucket that sank in 1896 and a helicopter that went down near there in 1986. "The longer you have to predict back in time, the bigger the rate of error grows. Once you get out into three or four weeks, it's really getting tough."
Ocean currents around the globe vary so widely that the only commonality they share is how easily they can hide objects lost at sea. Currents off Japan and South Africa are especially strong, as is the Gulf Stream, which flows north along the East Coast and Newfoundland then turns east toward Europe. Despite American scientists' familiarity with the Gulf Stream, Limeburner considers it one of the most challenging regions in the world to search.
"When the Gulf Stream leaves Cape Hatteras, it meanders around like an unattached garden hose," he says. "It has lots of variables and huge eddies—some are 100 miles across, and some are moving towards the east and some moving west. It's a complicated area."
The Indian Ocean is similarly complex, and the difficulties posed by its remoteness become more apparent with each passing day. The first search zone reached south into the southern Indian Ocean, toward the Antarctic Circumpolar Current, which flows east and circulates around the entire globe at the South Pole. Above it, flowing in the opposite direction, is the Indian Ocean Gyre, where the second and third search zones have been located.
In short, there is nothing simple or completely predictable about currents in this corner of the world. (See "Searching for the Missing Malaysian Jet at the Ends of the Earth.")
"You see eddies and whirlpools. Things can get wrapped around and go down plugholes. Things accumulate between different waters of different temperatures. Air temperatures change the wind. Everything in the ocean is a balance of forces," Griffin says.
Tracking the Buoys
Griffin's team has one thing going for it: research buoys that have been in use for more than 40 years. In the late 1970s, scientists from the National Oceanographic and Atmospheric Administration (NOAA), working with scientists from Australia and other countries, began dropping drift buoys regularly into the Atlantic, Pacific, and Indian Oceans for oceanographic and climate research. Tracked by satellites, the buoy data provides a coherent picture of ocean currents.
"We are very fortunate that quite a few buoys were deployed last year close to where the projected crash site is," he says. "We already know quite a bit about the trajectories that buoys take in this corner of the world—not enough to tell us where the debris has gone, but they tell us that our methods of testing these things are correct."
The trajectories showed how unpredictable the Indian Ocean can be. Griffin tracked one buoy as it appeared to move south in a straight line from western Australia to the southern Indian Ocean. When the satellite took a closer look, it revealed that the buoy was trapped in the center of a clockwise-rotating eddy and that the eddy was moving along in a straight line.
"It stayed inside that eddy for 15 months," he says. "That was one of those oh-my-god moments."
The search zones themselves are as uncertain as the waters being searched. They were drawn up using data from Inmarsat, the British company whose satellite picked up seven pings—the so-called handshake—from the Boeing 777 that seemed to indicate the jetliner flew on an arc across the southern Indian Ocean.
By contrast, the searchers looking for the Air France plane, which crashed in 2009, had the advantage of information unavailable to the flotilla now roaming the Indian Ocean. Investigators looking for the doomed Airbus knew the jet's airspeed, its last known location, and the time it fell from the sky.
Even so, the plane remained undiscovered until May 2011. In that case, searchers had unpredictable seas to contend with as well. The ocean currents where the plane went down divided and flowed in opposite directions: easterly in one area and westerly in another, just 60 miles away.
"So if you were looking halfway in-between, the plane debris could be in either of those two regions," Limeburner says.
Air France hired a team of scientists to model possible scenarios, just as the Australians have done now.
"Each one had their own model. It was like a talk show. Each has his own pet and thought his pet was the best," Limeburner says.
Even so, none of those models led to the wreckage. Finally, during the last 18 months of the search, Limeburner worked with another scientist and created a new set of models.
They deployed nine drifting buoys in an area around the last known position of the plane, and intentionally set them adrift at the same time of year as the crash had occurred.
"We did it a year later, with the understanding that it wasn't going to tell us what happened at the time, but would characterize the ocean for us at the time of the crash," Limeburner says. "We used data from those drift buoys to predict a path of debris."
They drew a circle around the area 20 to 30 miles (32 to 48 kilometers) in diameter and found the plane. It was 10 miles (16 kilometers) away from the last known position in flight.