Making Sense of Soil Via Big Data
By Zach Clemens
Ask any farmer, and they will explain the importance of soil. It is what moderates the long-term productivity of any field. Understanding the inherent properties of soil and how these vary across the landscape is vital to choosing management practices that maximize the longevity of the land’s value.
“We rely on soil for so many different things, the list can be overwhelming at times,” says Bradley Miller, assistant professor of agronomy. “Think about why the state of Iowa has the agricultural economy that it does. What makes this state unique? Largely, it’s the soil.”
A CYCLONE DOWN DEEP
Miller (‘00 environmental science, ‘06 MS soil science and water resources) is no stranger to Iowa soil, or Iowa State University. The Iowa native continued his education at Michigan State University with a Ph.D., then worked in Germany developing methods for digital soil mapping at the Leibniz Centre for Agricultural Landscape Research. He joined the ranks of his former mentors in 2015 when he accepted a faculty position at Iowa State.
“When I came out of high school, I was confused about what I really wanted to do. I received some nice mentorship from Nick Howell who I worked for at Reiman Gardens,” Miller says.
“Then one of my co-workers suggested that I meet with Lee Burras and the rest, as they say, is history.”
As an undergraduate, Miller worked as a research assistant learning about geographic information systems (GIS) and the need for better soil maps.
“When Bill Crumpton encouraged me to pursue a master’s, it was pivotal. Those mentorships really put me on the path to success,” Miller says.
Miller teaches geospatial technologies and digital soil mapping, while overseeing the department’s Geospatial Laboratory for Soil Informatics. He advises graduate students, and his research examines spatial variations in environmental processes that affect soil properties. He studies how those variations impact environmental quality and sustainable crop production.
UPDATING THE SOIL SURVEY
“We have been essentially mapping soil in Iowa the same way for almost 100 years,” Miller says.
The latest soil maps have only been slightly updated since they were created by Iowa State University, state and national conservation agencies and the counties of Iowa around three decades ago.
“The existing soil maps give a general idea of the soil resources. But, as we get into precision agriculture, a lot of farmers are using these maps to determine management zones within their fields,” Miller says. “The creators of these maps never intended them for this use, yet it is still the best data available for this purpose.”
Miller and his cadre of graduate students — Luis Bentancor, Meyer Bohn, Dustin Ehret, Caner Ferhatoglu and Emma Molburg — are taking a statistically based approach to accomplish finer spatial detail and reduce uncertainty in how soil properties are predicted.
“In some ways, we aren’t inventing the wheel. We are building on concepts from the past with the new algorithms we have in our toolbox,” Miller says. “The big difference is big data. We now have many sources of remotely sensed information, plus we have machine learning that helps us find much more complex patterns.”
Machine learning is a type of artificial intelligence in which computers create models based on large data sets. The different technologies Miller uses range from spectral data collected by satellites, to elevation data collected by lasers from airplanes.
The big questions Miller is tackling are, what is the best sampling design to capture the variation in the landscape; what are the best predictor variables to use; and what is the appropriate machine learning algorithm to find those complex patterns.
“Projects like Dr. Miller’s help us update and enhance tools used by about seven million people annually in a way that is more consistent over a larger area in less amount of time,” says David Hoover, director of the National Soil Survey Center in the U.S. Department of Agriculture Natural Resource Conservation Service.
“With digital soil mapping techniques, we can enhance those products or remap areas that have not been mapped before.”
A CERTIFIED STRONG FOUNDATION
With the help of Miller, the Department of Agronomy is offering a new soil science certificate. It allows undergraduates to receive official academic recognition in soil science, as opposed to advising themes or options within a major.
“Understanding soil is critical for addressing issues of food scarcity, infrastructure development, water management, climate change, biodiversity loss and human health,” Miller says.
The certificate will help students build a strong foundation in understanding soil systems and open more career opportunities requiring specialized education in soil science.
“Not only does completion of the certificate establish a student’s credentials for jobs at the federal level,” says agronomy professor Michael Thompson, who was part of the team who developed the soil certificate. “It also will be noted by other private and public employers to identify students with deeper knowledge and experience with soils than other graduates.”