South Dakota Mines Leads Research to Aid Understanding of Earth Systems with Big Data

A research team at South Dakota Mines led by Gokce Ustunisik, Ph.D., assistant professor of geology and geological engineering, has received $278,000 in grants from the National Science Foundation to help create a system that can compile experimental data from around the world on the elements that make up the earth's crust and mantle.

Igneous rocks and minerals are formed deep underground when magma cools and crystallizes. The specific minerals that come from cooling magma depends on many factors such as the makeup of the magma, the rates of cooling and how much pressure the magmas are under.

Scientists who study earth systems need to know the make-up and geochemistry of rocks to better understand how natural phenomena occur. An improved understanding of those phenomena can help geologists understand how the Earth's crust forms and how the forces acting on magmas drive plate tectonics. Understanding the connections between systems is equally important and can help us constrain critical information on the natural carbon flux out of the deep earth.

Researchers, like Ustunisik, use high temperature furnaces to recreate the conditions deep underground where superheated magma of various compositions cools and hardens. In her lab at South Dakota Mines, Ustunisik and her graduate student, Erica Cung, can melt rocks to help determine their composition. The data from experiments like this can give geologists an understanding to how rocks formed in a single location.

But to date, no one has been able to build a big picture with data on rock composition from multiple experiments from around the world. Ustunisik and Cung have compiled data from over 15,000 experiments as they develop an interface that provides investigators access to information that was previously distributed across over 700 different publications.

“Access to this database will both save time and taxpayer resources, and by compiling and sharing data, scientists can quickly see what has already been done so they are not reinventing the wheel with every experiment,” says Ustunisik.

Many challenges remain for the team; geochemistry of any given rock can be very complex. Additionally, the experiments on any of these rocks may have been done differently and they have been collected from many different parts of the world. The team must make sure they are comparing apples to apples in different data sets from different experiments.

“One of the challenges of big data is how to you enable the community to aid you in actually constructing the infrastructure,” says Roger Nielsen, Ph.D., a co-principal investigator with Ustunisik on the research.  “But without this dataset, no one can really get an idea of the big picture,” he adds.

Enhancing the big picture understanding of earth's deep history, how rocks form and the processes that govern things like plate tectonics, earthquakes and volcanoes is a major goal of the effort. “Compiling all this data will greatly improve computer modeling and make them more reliable,” says Ustunisik. She points out that scientists can take this global dataset to create a common language among researchers, do predictive modeling detailing the formation of and potential location of valuable minerals, and in the long run give us a much more detailed understanding of the composition of the earth itself.