March 7, 2018

TACC, Lamont Observatory Host One of the Largest Earth Sciences Data Collections in the Country

March 7, 2018 — The Texas Advanced Computing Center (TACC) at The University of Texas at Austin is partnering with the Lamont-Doherty Earth Observatory (LDEO) at Columbia University to host one of the largest data collections for Earth sciences of its type in the country. The data relates to the Ross Ice Shelf, a massive slab of floating ice that is about the same size as the country of France.

Over the past four years, researchers at LDEO have been flying over the frozen waters in the polar regions and collecting field data for the ROSETTA-Ice project, which studies the Ross Ice Shelf. The shelf is constantly fed by a flow of ice from glaciers draining from both the East and West Antarctic ice sheets. The field data includes crucial information on the shelf and the underlying tectonics of the Antarctic region.

Ice shelves, like icebergs, lie mainly below the waterline. This means that the majority of the shelf is not visible without the use of scientific instruments. Studying how the ice, ocean and underlying seabed interact will inform scientists of potential change in the ice shelf from projected climate change.

“The Ross Ice Shelf is of interest because it’s floating, allowing ocean water to move freely about beneath it and we have seen in other regions like this that they can become unstable and break up releasing ice from glaciers dammed up behind them into the ocean,” said Nick Frearson, a lead engineer on the ROSETTA-Ice project whose team designed the Icepod, the data collection system and sensor suite and the radar technology that probes the ice shelf.

“Warming ocean water is getting underneath the shelf that is a significant couple of degrees warmer than the surrounding water and can mean the difference between freezing and melting at the base of the ice,” he said. “The shelf acts like a large cork impeding the flow of incident glaciers and ice streams, and could have far-reaching effects if it changes significantly and releases more ice to flow from the land into the sea raising sea-levels globally in the process.”

Frearson says that the data being collected using scientific instruments — hundreds of terabytes in total— is unique. “We take data from a suite of instruments, all sampled synchronously, and bring them together to form a much bigger picture than if we just analyzed data from one instrument,” he said.

Up until now the sea floor under the shelf has only been mapped to a resolution of 50km using a combination of satellite gravity data and a land survey undertaken in the 1970’s. This is low enough to hide whole mountain and valley systems and was not detailed enough for oceanographers to accurately model ocean currents flowing under the shelf. With state-of-the-art radar; gravimeters, which measure gravity anomalies; a magnetometer that measures Earth’s magnetic anomalies; LIDAR, remote sensing of the surface with laser pulses; and high-resolution photogrammetry to map surface structures; ROSETTA has been able to map the interior and ocean floor of the shelf to much higher resolution.

“In the process, we have collected many 100’s of TB’s of data and needed a state-of-the-art solution to manage it. That’s where TACC comes in,” Frearson said.

To collect, process, analyze and store the data, Frearson and other colleagues at LDEO have been using the National Science Foundation-funded Extreme Science and Engineering Development Environment (XSEDE) allocations on resources such as Stampede2 and Ranch. Researchers also heavily relied on TACC’s Corral even though it is not an XSEDE resource. XSEDE is a single virtual system that scientists use to interactively share computing resources, data, and expertise.

Stampede2 is used for data processing; Corral for data storage; and Ranch tape storage for the long-term archiving of data.

“The speed of XSEDE and TACC resources is superior to our existing high-performance computers at Lamont,” said Lingling Dong, a computer software, and data engineer. “I have data from 2015 that processed in 50 hours using Lamont resources; however, it only needed three hours of processing time using XSEDE resources and a total of two hours for data-transfer across the network.”

When it comes to storage, LDEO had been storing the data locally. “But because of the size of it now, we couldn’t even back it all up,” Frearson said. “We had the original raw data that we brought back from the field as a backup but that was it. It’s a lot more data than we had previously been used to handling.”

Corral is the storage system of choice for the Polar Geophysics Group at LDEO and leads the way in the preservation and sharing of data for researchers. Corral enables data-centric science throughout the U.S. This storage and data management resource is designed and optimized to support large-scale collections and a collaborative research environment.

“It’s admirable that LDEO has planned for how much data they are going to generate and want to make sure that it’s available over a period of several years,” said Chris Jordan, manager of TACC’s Data Management and Collections group. “From the start, they wanted a way to both store hundreds of terabytes of data and to make it widely available on the web as it’s uploaded.”

In addition, LDEO is using some of the academic referencing mechanisms that make it easier to search for and locate data sets, similar to the way it is done with academic literature. “They chose to use DOI indexing on all of their data as they upload it so that people can find it more easily,” Jordan said. “The National Science Foundation promotes the citation of data sets in addition to the citation for publications.”

“We aren’t the only Science group having to cope with very large volumes of data and hope that the partnership that we have forged with TACC shows that it is possible to manage and disseminate this level of data in a cost-effective, user-friendly and easily accessible manner,” Frearson said. “This data will help people in the science community who are interested in the cryosphere, the Polar Regions in general, and the changes that are going on there. We hope that people across the globe, as well as institutions in the U.S., will benefit from this data set.”

Source: Faith Singer-Villalobos, TACC

Share This