Next-Generation Computing Platform Supported by $1.24M NSF Award
The Arctic climate system is undergoing rapid change, from rising sea and air temperatures to melting glaciers to the loss of permafrost and snow cover on land. As the Arctic thaws, maritime and commercial activities in the region are expanding, presenting new opportunities—as well as societal and cultural challenges. Habitats are eroding, and food systems are disrupted.
Scientists are urgently working to help understand and address these changes to the warmer, less stable “New Arctic.” The University and Maryland and Washington State University (WSU) are collaborating on a project, ICESpark, an open-source big data platform which unites the efforts of researchers in fields including geographical sciences, satellite remote sensing, artificial intelligence, computer science, cryospheric science, ecology and oceanography.
ICESpark was recently awarded a $1.24 million National Science Foundation EarthCube award. ICESpark is unique in its potential to adequately support and unite computing platforms accessible to scientists in these fields, breaking down silos and expanding capacity.
“ICESpark collaborators seek to improve understanding of changes underway in the Arctic region, with tools that will be applicable across all of the geosciences,” said Assistant Professor of Geospatial Information Science Yiqun Xie, the Principal Investigator for the UMD award.
The project started in 2020 when Xie and Assistant Professor Jia Yu of WSU (Principal Investigator for the WSU award), sought to combine their expertise on spatial big data platform and data science to build a next-generation spatial computing system with full-stack capabilities from data management to advanced analytics.
“Soon, we realized it will be far more interesting for us to step outside the pure technical mindset and design the platform around important science questions which may have more significant societal impacts and benefits,” Xie said.
Gradually, collaborators began to join the project, and the team is focused on researching the New Arctic and related impacts from multidisciplinary angles and using a variety of methods.
The work is not only innovative, but is addressing a timely topic—while the Arctic region seems far away to many people, changes in the region are having an impact close to home for many humans.
“We’re interested in learning more about changes underway in the polar regions because mounting evidence suggests that changes in Arctic sea ice are impacting mid-latitude weather. For example, a weakening of the poleward temperature gradient appears to amplify meanders in the jet stream, resulting in more persistent weather patterns and a higher frequency of extreme weather events—heavier rainfall and extreme winter storms—at mid-latitudes,” said Associate Professor of Geographical Sciences Sinead Farrell, Co-PI. Farrell also holds an appointment in the Department of Atmospheric and Oceanic Science in the College of Computer, Mathematical, and Natural Sciences (CMNS).
Among the first issues that ICESpark will tackle is to co-locate satellite data with New Arctic measurements made in the field, including information gathered by autonomous buoys or by researchers at sea or based at ice camps.
This type of work has traditionally been time-consuming and difficult, because of the sheer volume of satellite data, which is hard to match with field information.
“It’s an important job, though, because the field data allows us to evaluate the quality and accuracy of the satellite measurements. This in turn gives us confidence in the remote sensing data when we look at changes across vast areas like the global oceans or the polar regions,” Farrell said.
ICESpark will be an especially invaluable tool for oceanographers, said Assistant Professor Jacob Wenegrat of the Department of Atmospheric and Oceanic Science, Co-PI.
“Oceanography over the last decades has gone from a data-sparse field, where observations require people to go to sea in ships, to an increasingly data-rich field. Much of this has come about due to satellite remote sensing and the rise of autonomous platforms for ocean sensing,” Wenegrat said. “The challenge now becomes, ‘how do we extract as much scientific insight as possible from this data?’ ICESpark offers the ability for more rapid discovery, and in particular the ability to rapidly combine data across a variety of data sources—satellites, autonomous floats, drifters. ICESpark offers tools that let scientists focus on science, finding new creative ways to analyze and understand big data about our planet.”
In the short term, the ICESpark team hopes to improve the tool’s flexibility, scalability and efficiency in handling geospatial big datasets from multiple sources.
“We want to perform a variety of tasks, ranging from data discovery to geo-feature detection to geo-pattern mining,” Xie said. “Our geoscience experts will also be deeply involved in system evaluation to make sure the design and development align with real needs.”
In the long term, the team hopes that ICESpark will address urgent environmental questions.
“We will measure success by the number of groups worldwide using ICESpark to analyze not only polar remote sensing data sets, but also for applications in ecology and oceanography. One of the major goals of the NSF EarthCube program is to provide benefits across the geoscience community and build lasting collaborations. Our team will seek to work closely with existing NSF EarthCube projects to optimize the utility of ICESpark in addressing community needs,” Farrell said.
ICESpark’s team also includes Professor and Research Director George Hurtt of the Department of Geographical Sciences, Co-PI; Senior Faculty Specialist Kyle A. Duncan of the Earth System Science Interdisciplinary Center in CMNS, Faculty Associate; and Collaborator Twila Moon, PI of another EarthCube grant, “QGreenland,” at the University of Colorado and National Snow and Ice Data Center.
Photo via iStock.