Supercomputing comes to Ag

(University of Minnesota Supercomputing Experts Kevin Silverstein and Naomi Hospodarsky will help the Soil Health Partnership bring powerful analysis to its data.)

Written by Jonathan Eisenthal

The power of supercomputing has helped bring new levels of accuracy to many areas, including weather forecasts as one example.

A National Corn Growers Association initiative, the Soil Health Partnership (SHP) has begun working with the University of Minnesota’s interdisciplinary supercomputing group called GEMS to bring that predictive power to conservation agriculture.

With sound data mixed with the power of supercomputing, a farmer would be able to choose the right mix of methods and plant materials to get the best possible crop yields while meeting their environmental goals. This future is still years off, but these partners are hopeful.

“The analytical power we can bring to the Soil Health Partnership data is immense,” said Kevin Silverstein, the operations manager of GEMS, which stands for Genetics X Environment X Management X Socioeconomics.

But before analysis, GEMS has to dive into SHP’s data.

Silverstein said it comes down to making sure that units of measurement match, and that each datum comes with the necessary meta-data attached to it. In the case of corn yield data, it’s critical to know the exact place, the day and date of harvest, planting date, climate conditions over the entire growing season—there’s a volume of meta-data that has to match up with how many bushels came off a particular acre of land.

“(GEMS has) some really cool tools they have built around data cleaning and data interoperability, with data being collected from all over the world,” said SHP Senior Scientist Maria Bowman. “Even outside of our own data set, there are different types of data that we want to incorporate to answer questions that we may have: How do our yields compare to USDA yields? Do soil health practices have an impact on yield variability?

GEMS has already successfully cleaned the corn breeding data developed by two dozen land grant universities, in a program called Genomes to Fields. Professors Candace Hirsch and Nathan Springer at the University of Minnesota are contributing scientists in the project.

Silverstein estimates this process will take about a year, and then SHP can begin asking questions of the data.

“The challenge is that we’re trying to quantify the impacts on a living system,” said John Mesko, senior director of SHP. “People want to know, if I am planting cover crops, what is going to be the impact in terms of dollars per acre, and can I count on that, and can I take that to my banker?”

Right now, SHP has reached a small group of ‘early adopters,’ including eight farmers in Minnesota. Mesko and Bowman hope that the power unleashed by supercomputing will bring a new set of growers into the fold.

“These are people who are more than willing to make changes if it can be demonstrated that they are valuable. A lot of times those producers are among the more influential farmers—they are often times larger, and they are often times more well known in their communities. We think this is a way to drive further adoption (of soil health practices) as we move forward.”

Learn more about efforts from the Soil Health Partnership at soilhealthpartnership.org.

Did you like this article?

Share this post with your friends!