Studying climate change in the cloud

Boston University’s Hariri Institute for Computing and NASA’s Jet Propulsion Laboratory join forces to create a cloud computing platform for global change scientists.

BY GINA MANTICA

For media inquiries, please email gmantica@bu.edu.

Climate change might be partly to blame for the severe winter weather in Texas. To create effective policies that protect the Earth’s climate, researchers need to bett­er understand how the Earth’s land cover, or the physical makeup of the land, changes over time. But researchers want to access, analyze, and process large amounts of complex land cover data more easily.

Researchers from Boston University’s (BU) Hariri Institute for Computing and Center for Remote Sensing have teamed up with scientists at NASA’s Jet Propulsion Laboratory (JPL) to create an accessible cloud-based system for storing and modeling very large and complex land cover data sets. The BU Hariri Institute for Computing recently granted Mark Friedl and Luis Carvalho funding to support their collaboration with NASA JPL’s Thomas Huang and Dave Schimel.

Earth observation satellites, like NASA’s Surface Biology and Geology Mission, provide researchers with petabytes of sophisticated data on Earth’s land and its properties. These satellites not only capture information about different types of land cover, they also measure changes in critical environmental properties such as the density of forests, the amount of nitrogen in forest canopies, the different types of plants on the ground, and even the composition of the rocks.

The information collected by satellites through a process known as remote sensing helps researchers study the local effects of climate change on ecosystems all over the world. Researchers can use remote sensing data to model how much carbon terrestrial ecosystems can sequester and how much carbon is being released into the atmosphere as greenhouse gases.

But scientists need a system to efficiently analyze these enormous and complex remote sensing data sets using high performance computing, where many computer servers work together to finish a task by dividing up the computational load. “We are literally drowning in data and we don’t have the ability to analyze and process the data in the way that we’d like,” said Friedl, “We really need cloud computing.”

Cloud computing, or using very large networks of remote servers hosted on the internet for data management and analysis, provides plentiful storage space and computing power for researchers to analyze remote sensing data. To support this effort, the BU Hariri Institute for Computing and NASA JPL teams plan to install an open-source Science Data Analytics Platform (SDAP) on the Mass Open Cloud. The researchers will also install land cover datasets and develop algorithms for land cover analyses on the SDAP, leveraging storage available through the Northeast Storage Exchange (NESE).

“Cloud computing is the future for NASA’s upcoming big data systems, and we’re excited to partner with BU to develop new tools to make the cloud environment ideal for academic and applied users of our next generation of high resolution and highly informative land surface and aquatic observations,” said Schimel.

The cloud computing platform will catalyze global change research by providing a cost-effective way for scientists to run custom land cover analysis algorithms. The researchers hope to scale up their platform to eventually run land cover analyses across the globe. “The more we can get good tools into the hands of good scientists, the more we can remove barriers to them doing their science,” said Friedl, “Our ultimate goal is to provide the best possible science to inform public policy.”

Interested in learning more about the transformational science happening at the Hariri Institute? Sign up for our newsletter here.