ORNL researchers are demonstrating how integrated automation, robotics, imaging, sensors, and real-time edge computing capabilities can accelerate scientific discoveries in field experiments, providing insights into strategies for enhancing soil carbon storage.
Transforming the current pace and scale of measurements belowground is critical to address knowledge gaps about how the spatial and temporal variability in soils affects carbon and nutrient cycling. To meet the grand challenge of sequestering atmospheric carbon dioxide in soils at landscape-scales, a deeper understanding and more data are needed.
These data will inform predictive models to speed research and development of crops with higher carbon storage capacity above and belowground — crops that are essential to building a sustainable bio-based carbon-neutral economy.
Demonstrating a new, automated approach to field science
As part of this project, researchers designed and are building small, all-terrain robots that can dig and collect soil cores – columns of soil that scientists study to understand how carbon, nutrients, and other factors vary belowground in different locations and at various depths. Onboard sensors send data about the soil samples to a computing station at the field’s edge in real-time.
Scientists are ground-truthing the data the system collects by comparing it to samples secured through more traditional, manual methods. The field is planted with 200 variants of poplar, a fast-growing bioenergy feedstock, and researchers have been evaluating soil, root, and leaf samples over the past year.
ORNL-developed technology commercialized by Campbell Scientific to monitor plant productivity and health is also deployed at the site. The technology, known as Fluorescence Auto-Measurement Equipment, or FAME, measures solar-induced fluorescence, which is emitted by plants during photosynthesis as they convert sunlight into chemical energy.
The team developed the necessary cloud communication and data handling workflows to enable the edge computing station to collect the data from these aerial and belowground sensors. In a future phase of the project, researchers could write the algorithms necessary for real-time automated assessment of the sample data to create a self-guided system that directs and executes on sampling protocols to inform carbon sequestration research.
This proof-of-principle project demonstrates how automation can speed scientific discoveries in the field. It is the first step in producing a lab-to-field plant performance prediction capability to accelerate development of better crops in support of a sustainable bioeconomy.
