A team of ocean robots deployed in January 2018 have, over the past year, been the first self-guided ocean robots to successfully travel under an ice sheet and return to report long-term observations.

Beyond mere survival, the robotic mission — a partnership between the 91̽»¨’s College of the Environment, the 91̽»¨Applied Physics Laboratory, the Lamont-Doherty Earth Observatory of Columbia University, the Korean Polar Research Institute and Paul G. Allen Family Foundation — has ventured 18 times under the ice shelf, repeatedly reaching more than 40 kilometers (25 miles) into the cavity, among the farthest trips yet into this treacherous environment.

“This is the first time any of the modern, long-endurance platforms have made sustained measurements under an ice shelf,” said , a 91̽»¨professor of oceanography and member of the Applied Physics Laboratory. “We made extensive measurements inside the cavity. Gliders were able to navigate at will to survey the cavity interior, while floats rode ocean currents to access the cavity interior.

“It’s a major step forward,” Lee added. “This is the first time we’ve been able to maintain a persistent presence over the span of an entire year.”

The project funded by Paul G. Allen Family Foundation seeks to demonstrate the technology and gather more data from the underside of ice shelves that are buttressing the much larger ice sheets. Direct observations of how warmer seawater interacts with the underside of ice shelves would improve models of ice sheet dynamics in Antarctica and Greenland, which hold the biggest unknowns for global sea level rise.

“Some ice sheets terminate in large ice shelves that float out over the ocean, and those act as a buttress,” Lee said. “If the ice shelves collapse or weaken, due to oceanic melting, for example, the ice sheets behind them may accelerate toward the sea, increasing the rate of sea level rise.”

“Most of the uncertainty in global sea level rise predictions for decades to centuries is from ice sheets, which could contribute from 1 foot to as much as 6 feet by 2100,” said , a research professor of oceanography at the Lamont-Doherty Earth Observatory. “A key driver is interaction with the ocean heat and these new tools open tantalizing perspectives to improve on current understanding.”

The mission set out in late 2017 to test a new approach for gathering data under an ice shelf, and on Jan. 24, 2018, devices were dropped from the Korean icebreaker R/V Araon. This week, two self-navigating Seagliders reached the milestone of one year of continuous operation around and under the ice shelf.

Robot submarines operated by the British Antarctic Survey, known as Autosub3 and , successfully completed 24- to voyages in 2009, 2014 and 2018. These missions surveyed similar distances into the cavity but sampled over shorter periods due to the need for a ship support.

By contrast, the U.S.-based team’s technology features smaller, lighter devices that can operate on their own for more than a year without any ship support. The group’s experimental technique first moored three acoustic beacons to the seafloor to allow navigation under the ice shelf. It then sent three Seagliders, swimming robots developed and built at the UW, to use preprogrammed navigation systems to travel under the ice shelf to collect data.

The mission also deployed four UW-developed EM-APEX floating instruments that drift with the currents at preselected depths above the bottom, or below the top of the cavity, while periodically bobbing up and down to collect more data. All four of these drifting instruments successfully traveled deep under the ice shelf with the heavier, saltier water near the seafloor. Three were flushed out with fresh meltwater near the top of the ice cavity about six to eight weeks later. One float remained under for much longer, only to reappear Jan. 5.

During the past year, the fleet of robots has reached several milestones:

• A Seaglider reached a maximum distance of 50 kilometers (31 miles) from the edge beneath Dotson Ice Shelf in West Antarctica;
• The Seagliders made a total of 18 trips into the cavity, with the longest trip totaling 140 kilometers (87 miles) of travel under the shelf;
• The Seagliders also made 30 surveys along the face of the ice shelf;
• After one year, two out of three Seagliders are reporting back;
• In the current Southern Hemisphere summer, one of the Seagliders has gone back under the ice shelf and has completed two roughly 100-kilometer (62-mile) journeys;
• Another Seaglider will begin its second year of sampling at the face of the ice shelf;
• Three drifting floats journeyed under the Dotson Ice Shelf and back out in early 2018;
• After 11 months under the ice, the fourth float reported home in mid-January 2019 close to the neighboring Crosson Ice Shelf.

Researchers are now analyzing the data for future publication, to better understand how seawater interacts with the ice shelves and improve models of ice sheet behavior.

Four months of data show three Seagliders dropped from the ship in late January, then traveling toward the Dotson Ice Shelf (white). Two Seagliders (pink and orange) venture under the ice sheet in summer, while a third (yellow) samples along the face. The gliders then spend the colder months sampling along the ice sheet’s edge. Meanwhile, the drifting floats are dropped closer to the ice edge in late February. The teal tracks show how they drift under the ice sheet and then get flushed out in late March. A fourth float drifted to the right of this image, reaching a neighboring ice sheet.

Other members of the team are , a 91̽»¨assistant professor of Earth and space sciences who is currently in Antarctica on a separate project; , and at the Applied Physics Laboratory; and the Korean Polar Research Institute, or KOPRI.

###

For more information on the Seaglider component, contact Lee at craiglee@uw.edu or 206-685-7656; on the drifting floats, contact Girton at girton@uw.edu; and for more general questions, contact Dutrieux at pierred@ldeo.columbia.edu or 845-365-8393.

Images and video are available for download at .

Nicknamed the “,” Thwaites Glacier already is contributing to rising seas; if it collapsed it would raise global sea level by about three feet. The glacier may also act as a linchpin on the whole West Antarctic Ice Sheet, which could raise sea level by much more.

91̽»¨ glaciologists will participate in one of the eight projects funded through the to better understand the glacier and predict what it will do next.

“About 100 scientists are involved in this initiative, which is the largest Antarctic deep-field effort in 70 years,” said , assistant professor of Earth and space sciences and the UW’s principal investigator on the project. “This is one of the largest deep-field efforts ever attempted in West Antarctica, and is on a scale neither the U.S. nor the U.K. — or anyone else — could accomplish alone.”

The 91̽»¨is participating in the , or GHOST project, that will collect on-the-ground data to see the details of the glacier’s internal structure and better map the surface underneath.

The information they collect will provide better data to feed into the computer models that scientists are using to forecast the future of Earth’s climate.

“It’s unlikely that the ice-sheet modelers have the big story wrong,” Christianson said. “But if you’re looking for shorter-term estimates, having detailed information about the bed conditions or how ice is flowing over a ridge can be useful for understanding how the glacier might behave over the next few decades or century.”

That’s the kind of timeline that would be useful on the ground. A 2014 found that Thwaites Glacier would likely collapse within 200 to 1,000 years, though other estimates also exist. More data could provide a firmer timeline.

The GHOST team plans to spend two 60-day seasons in the field during the Antarctic summer, beginning in late 2019 and late 2020. An initial trip later this year will install equipment and fuel caches and survey potential base camps.

During the field campaign, the team will begin near the glacier’s coastal terminus, where the most rapid changes are occurring. Christianson, 91̽»¨postdoctoral researcher and a graduate student will conduct the scans. The 91̽»¨team plans to travel by snowmobile, surveying about 50 km (31 miles) of the glacier per day. Every week or so the entire GHOST project’s science team will move and set up a new camp, gradually working its way uphill toward the ice sheet’s interior.

The 91̽»¨researchers will use two different radars to map individual layers of snow and ice, and the underlying bedrock. The first technology, ground-penetrating radar, has been used by Christianson’s group for years and can penetrate through more than 2 miles of ice. His team’s data will complement airborne surveys being done by other teams.

“To map out a glacier’s internal structure, it’s advantageous to have a ground-based survey,” Christianson said. “If you want to recover very steeply-sloped layers of ice where the ice has folded, you can acquire higher-quality data if you’re driving on the ground. These layers tell us about the past flow structure of the ice, as well as the current deformation: where it’s bending in response to a sticky edge, or an area where there’s a lot of friction at the base.”

The other method the 91̽»¨team will use is a newer tool developed by the British Antarctic Survey that sends similar waves into the glacier but is designed for determining relative rates of change between layers in the ice detected by the radar. This technique allows detection of changes in glacier internal structure with a scale of just fractions of an inch, rather than several feet. Christianson hopes to repeat these scans a few weeks apart to see the details of Thwaites Glacier’s movement during a single summer.

“By taking the difference of two measurements we can see the ice deform and study how it flows in near-real time,” Christianson said. “This technique has been used in glaciology for about the last five years, but we’ll be doing it on a new scale for this project.”

The GHOST project is led by Sridhar Anandakrishnan at Pennsylvania State University and Andy Smith at the British Antarctic Survey, and involves scientists from several other U.S. and U.K. institutions.

Watch a video featuring leaders of the overall Thwaites Glacier project:

###

For more information on the 91̽»¨effort, contact Christianson at knut@uw.edu. More information on the larger project is at .

A new partnership between the 91̽»¨’s College of the Environment, the 91̽»¨Applied Physics Laboratory and Paul G. Allen Philanthropies will use a robotic network to observe the conditions beneath a floating Antarctic ice shelf.

Ice shelves act as buttresses that restrain the flow of inland ice into the sea, which under a warmer climate could trigger many feet of global sea level rise, on a timeline that is largely unknown. Observations in the water-filled caves under ice shelves could help explain how warmer seawater interacts with the glacier’s underbelly.

The team members performed a final test Nov. 6 in Puget Sound before the instruments are deployed in the Southern Ocean from a Korean research ship, the R/V Araon, that departs from New Zealand in mid-December.

“A project as experimental as this one would be impossible without the support of Paul G. Allen Philanthropies,” said , a 91̽»¨professor of oceanography and oceanographer at the 91̽»¨Applied Physics Laboratory. “This is a high-risk, proof-of-concept test of using robotic technology in a very risky marine environment.”

The ice shelf is the floating portion of a glacier that extends seaward from inland ice, which rests on bedrock. Most of Antarctica does not yet show significant surface melt, but scientists think melt is happening at the glacier’s underbelly, where relatively warm ocean water meets its underside. What is learned with this new data will help scientists better understand the stability of these ice shelves and help make predictions about sea level rise.

“This is one of a series of philanthropic investments by Paul Allen to improve our understanding of how the Earth is changing and how it’s being impacted by climate change,” said , director of climate and energy for Paul G. Allen Philanthropies.

91̽»¨oceanographers invented the Seaglider in the mid-1990s, with support from the National Science Foundation, and still build research models of the torpedo-shaped ocean drone. 91̽»¨researchers adapted the Seaglider for operating under ice, and have been using it to sample below Arctic sea ice since 2008. In 2014, Lee used a Seaglider and other technology in the Arctic Ocean to .

This new project will deploy a similar robotic network in the Southern Hemisphere. The environment is more challenging because the instruments must venture into the ocean cavities formed by ice shelves, which are very complex, but largely unknown, environments.

“We have almost no information about the area where the glacier is floating on top of the ocean,” said glaciologist , a 91̽»¨assistant professor of Earth and space sciences. “The ice is 300 to 500 meters (1/5 to 1/3 of a mile) thick. There’s no light penetrating, it’s impossible to communicate with any instruments, and this environment is extremely hard on equipment — picture big crevasses, rushing water and jagged ice.”

This effort included figuring out how to develop gliders that can get in and out from the ice sheet’s edge without being crushed by moving ice, swept away by fast-flowing water or trapped in the complex of ridges and crevasses on the ice shelf’s underside.

This year’s test also will use complementary technology designed by , an oceanographer at the 91̽»¨Applied Physics Laboratory, which drifts with the currents while moving up and down gathering data.

The team has devised new navigation algorithms for the Seaglider and tested them in simulations to make sure the instrument can navigate and return safely. The plan is for the gliders to initially travel in and out of a cave several times a day in summer, surfacing between each trip to beam data back to shore.

Once the ocean surface freezes for the Southern Hemisphere winter, the robots will continue to take measurements on their own, and will beam data back only when they emerge months later in the spring.

“We’ve never been able to get really deep into an ice cave, where the floating ice shelf meets the seafloor,” Christianson said. “If we can do that, we’ll be able to collect tons of new data. We often don’t even know what the topography of the seafloor is like beneath the shelf, which affects water flow, temperature and other factors that control the melting rate.”

Team member , a glaciologist at Columbia University’s Lamont-Doherty Earth Observatory, has used other technologies to gather more limited observations below ice shelves. He and , an oceanographer at the 91̽»¨Applied Physics Laboratory, will travel to Antarctica in December for the first deployments under Pine Island Glacier, if conditions allow, or another nearby extension of the West Antarctic Ice Sheet. They plan to deploy three gliders and four floats and leave them down for a period of about a year.

The Korean Polar Research Institute (KOPRI) is also partnering for this mission. KOPRI will provide field support for the deployments from its ice breaking research vessel Araon, will conduct complementary measurements from the ship and will collaborate on the subsequent analysis of the resulting data.

###

For more information, contact Lee at craiglee@uw.edu or Christianson at knut@uw.edu.