A Robot Rides Out Sandy, Gathering Data to Predict Storm Intensity
As millions of East Coast residents scrambled to get out of the way of “superstorm” Sandy this week, Liquid Robotics was steering one of its seagoing robots directly into the storm’s path.
The Sunnyvale, CA-based company says its robot, a “Wave Glider” model named Mercury, rode out the storm unscathed, and was able to transmit data on water temperature, barometric pressure, and wind speed back to controllers in near-real time.
Though Mercury wasn’t originally equipped to take on a hurricane, its adventure showed how fleets of autonomous robots could help weather modelers make better predictions about the intensity of hurricanes and tropical storms, says Edward Lu, Liquid Robotics’ chief of innovative applications.
“The prediction cones that you are seeing for hurricanes are getting pretty good, because we understand the upper-level winds that steer them,” says Lu, a former NASA astronaut who has logged more than 200 days in Earth orbit. “Seventy-two hours ahead of landfall, we have a pretty good idea of where a hurricane is going. What we do not have is the ability to say how strong that hurricane is going to be when it hits shore. That’s information residents and decision makers need.”
Wave Glider robots, which consist of a surfboard-like surface platform tethered to an underwater “glider” that provides passive propulsion, are designed for year-long, autonomous missions to gather data that can be used for security and defense, climate monitoring, and energy exploration. Thanks to instruments on the glider and the tether, Wave Gliders also have a unique ability to measure water temperatures in a column extending several meters below the ocean surface, Lu says. Using that information, meteorologists can infer how much heat has been sucked up by a storm system, and gauge its power as it approaches land.
“That’s data you can’t quite get from a satellite or an aircraft, and you don’t want to send a ship with people aboard into a hurricane,” Lu says. “We’ve shown, A, that the Wave Gliders are pretty darn robust and that we can steer them in the direction we want, and B, that we can get the data in real time.”
During Sandy, the Mercury craft measured a drastic plunge in barometric pressure—from around 1,000 millibars to a low of 946 millibars as the storm approached land. (The lower the pressure, the more powerful the storm; Hurricane Wilma in 2005 set a record low of 882 millibars.) Wind speeds were so high they exceeded the 70-knot limit of the craft’s small, commercial-grade anemometer, Lu says.
This isn’t the first time a Liquid Robotics craft has ridden out a big storm. “Over the last year and a half we’ve had many Wave Gliders go through hurricanes, but recently we’ve begun doing it in purpose,” Lu says. The company is funding the effort on its own, but it’s sharing the data with the National Oceanographic and Atmospheric Administration and the National Weather Service, in the hope that the federal agencies will eventually incorporate it into their prediction models.
“We thought it was important enough project that we could move much faster if we just funded it ourselves,” says Lu. “Ideally we’d like to do this in a funded way, because then it would be on a much larger scale.”
Lu envisions a future where whole fleets of Wave Gliders are patrolling areas like the Eastern Seaboard, the Gulf Coast, and even the South Pacific, helping to inform citizens of approaching storms, the same way NOAA’s Pacific Tsunami Warning System alerts coastal populations to threats from seismically generated ocean swells.
Data on ocean-water temperature at various depths is critical in deciding how to prepare for a hurricane, Lu says. These storms are basically giant steam engines heated by warm water; the more heat is transferred from the ocean to the atmosphere, the more severe a storm will likely become. “It’s not just the upper millimeter of water that counts,” says Lu. “As soon as a hurricane starts to go over an area, it stirs up the water, which is a different temperature below, and it quickly gets mixed up and the water cools down. If you know how much it’s cooled down, you know how much energy has gone into the hurricane.”
Mercury was not one of the Wave Gliders designed to study hurricanes, but it happened to be in a position—about 100 miles offshore from Toms River, NJ—that allowed Liquid Robotics controllers to steer it into the path of Sandy. The craft collected data throughout the storm, transmitting it back to Liquid Robotics every 30 minutes via a radio connection to the Iridium satellite network. “You can see the giant drop in pressure. You can also see the wind picking up and changing direction in huge gusts. It’s difficult to say that something is significant from one data point, but what we have shown that we have the capacity to measure what we want to know.”
While the current hurricane season is (hopefully) winding down, Liquid Robotics plans to study more storms next season, and to send robots into Southern Hemisphere typhoons and cyclones. “Next year we would like to increase the coverage and work closely with NOAA to incorporate this data into the hurricane prediction models,” Lu says.
Lu, an electrical engineer by training, joined Liquid Robotics in 2011 after several years as leader of the Advanced Projects Group at Google, overseeing imaging for Google Maps and Google Earth, among other projects. Before Google, Lu spent 12 years with NASA, where he flew on two Space Shuttle missions and one Russian Soyuz launch and spent six months aboard the International Space Station.
“I’ve spent a lot of time looking down into hurricanes from above, and they are unbelievable,” Lu says. “What is amazing to me is that you can run one of these little robots through them and they come out the other end unscathed and can transmit data right through the middle of the storm. This is kind of game-changing, because we need to learn better what threats are coming our way.”