New AI Method Advances Prediction of Brazil’s National Soybean Yield
URBANA, Ill. [26th Nov. 2025] — A new study from the Agroecosystem Sustainability Center (ASC) of the University of Illinois Urbana-Champaign provides one of the most comprehensive explanation to date of how tile drainage, a common agricultural practice, enhances the functioning of agricultural landscapes. Although tile drainage has been widely studied as an important form of agricultural infrastructure, the new study built a comprehensive framework to explain why tile drainage is so effective across a wide range of outcomes. The study pinpoints soil oxygen dynamics as the critical, hidden mediator that is pivotal for drainage impacts on crop growth, soil health, and crop resilience.
The study recently published in Hydrology and Earth System Sciences demonstrates that the benefits of tile drainage extend far beyond simply removing excess water from fields. Tile drainage fundamentally alters soil hydrology by reducing soil water content, which then enhances soil oxygenation. These hydrological impacts have complex effects on soil biogeochemistry and plant biology. For example, the improved aerobic condition alleviates crop oxygen stress during wet springs, which, in turn, promotes early crop root growth. The increase in oxygen availability also stimulates microbial activity, which accelerates the decomposition of organic matter and nutrient cycling.
The research team used an advanced, process-based model called ecosys that is uniquely capable of simulating the physics of soil oxygen movement and crop oxygen uptake. After validating the model against multi-year field data, the researchers ran simulations comparing drained and undrained conditions to understand the full set of ecological changes following drainage.
According to Kaiyu Guan, the project’s principal investigator and a Levenick Endowed Professor and Director of the Agroecosystem Sustainability Center at the University of Illinois, “Previous models that do not explicitly simulate soil oxygen dynamics fail to capture the true impacts of tile drainage, making it hard to accurately predict agriculture outcomes such as yield, nutrient availability, and nutrient leaching. These outcomes and the underlying processes must be understood through the lens of soil oxygen.”
“Most models use soil water as a simple proxy for oxygen stress, but without explaining the underlying physical mechanisms that improve crop outcomes. Our work focuses on the central role of oxygen. Its availability depends on the balance between supply (oxygen diffusion through soil) and demand from roots and microbes. This allows us to provide a holistic view of how tile drainage impacts the broader agroecosystem,” said lead author Zewei Ma, a PhD student at the University of Illinois under Prof. Guan.
The authors note that while conventional tile drainage provides clear agronomic benefits, emerging practices such as controlled drainage offer a more comprehensive approach that help mitigate potential water-quality trade-offs. By adjusting outflow during different parts of the season, controlled drainage can retain more water and nutrients in the field while still improving soil oxygen dynamics when crops are most vulnerable to oxygen stress.
The key findings of the study include:
- Soil oxygen as the key controlling factor: By removing excess water, tile drainage actively oxygenates the soil. The improved aerobic environment results in a range of soil health benefits.
- Stronger root development: Improved soil oxygen levels during wet springs alleviate stress on crop roots, allowing them to respire and grow more efficiently. This leads to the development of denser and deeper root systems.
- Accelerated nutrient cycling: Higher oxygen levels stimulate microbial activity, which accelerates organic matter turnover and increases nutrient availability for crops.
- Water quality trade-off: The study also confirmed that tile drainage can lead to increased leaching of soil nitrogen into waterways, highlighting the need for paired conservation practices to protect water quality.
“This research provides a mechanistic understanding of why drainage benefits crops. It goes far beyond water management; it improves the conditions for microbes and crop roots in the soil. By reducing oxygen stress in the root zone, the plants have greater ability to establish a resilient foundation for the entire growing season,” said co-author Bin Peng, an assistant professor on agricultural water management and water quality at the Department of Crop Sciences at the University of Illinois.
Professor Guan emphasized the broader implications, saying: “As we face a future with more climate extremes, strategic water management is essential for food security. This study gives us a powerful predictive tool to assess where and how tile drainage can best serve as an adaptation strategy, not just for increased yield, but for greater long-term yield stability.”
The researchers hope their findings will inform farmers, agricultural advisors, and policymakers on the multifaceted value of drainage management and the importance of integrating it with other practices to ensure both productivity and environmental sustainability.
The paper, “Soil oxygen dynamics: a key mediator of tile drainage impacts on coupled hydrological, biogeochemical, and crop systems,” is published in Hydrology and Earth System Sciences [DOI: 10.5194/hess-29-6393-2025]. The work was supported by the National Science Foundation, the U.S. Department of Agriculture, the Foundation for Food & Agriculture Research, and the U.S. Department of Energy.
For more information, contact:
Kaiyu Guan, Professor
Department of Natural Resources and Environmental Sciences
University of Illinois Urbana-Champaign
kaiyug@illinois.edu
Bin Peng, Assistant Professor
Department of Crop Sciences
University of Illinois Urbana-Champaign
binpeng@illinois.edu
Author: Lesly Goh. On September 5, 2025, I had the privilege of speaking at the FAO Inter-Regional Digital Agriculture Solutions Forum (IDASF 2025) in Bangkok. This global event brought together policymakers, researchers, and innovators to explore how digital technologies can accelerate rural transformation and bridge the digital divide in agrifood systems.
The Role of Academia in Digital Inclusion
My panel, “Role of academia in powering digital innovations and enhancing digital inclusion,” focused on how universities and research institutions can drive impactful change through cross-sector partnerships. Academia plays a critical role in translating cutting-edge research into practical solutions that empower rural communities, strengthen food security, and promote sustainability.
The Vision Behind SIGMA
The inspiration for my talk came from my experience at the Rockefeller Foundation’s Bellagio Residency in September 2023. That experience helped me crystallize the concept of developing a multi-stakeholder rice ecosystem. Over the past year, I’ve been collaborating with top scientists at the University of Illinois and University of Arkansas to operationalize this concept, moving beyond research to translate technology directly into local communities in Southeast Asia.
Introducing SIGMA: A Global Model for Agri-Food Systems
This effort led to the formation of the Southeast Asia Innovation Alliance for a Global Model of Agri-Food Systems (SIGMA). Co-led by the University of Illinois and the National University of Singapore (NUS), SIGMA is an academic research initiative that digitally captures plant growing systems, starting with rice. We leverage advanced sensing, modeling technology, and rich data to understand how crop systems respond to different conditions. SIGMA is designed to transform academic innovation into real-world solutions for economic development, improve farmers’ livelihoods, and promote environmental sustainability in Southeast Asia. Learn more about SIGMA: https://asc.illinois.edu/sigma/
A Solution for the Digital Divide: Our SYMFONI MRV Technology
During my presentation, I discussed the limitations of current Measurement, Reporting, and Verification (MRV) technologies, many of which require extensive data and are too costly or complex for smallholder farmers. These systems often work well for large, consolidated farms with strong digital infrastructure but pose a significant burden on smallholders in developing countries.
To address this, the SIGMA team developed a system-of-systems approach called SYMFONI, which is specifically designed to overcome these challenges. SYMFONI combines process-based modeling, ground data collection, remote sensing, and AI-empowered model-data fusion. This approach minimizes the need for extensive manual data input from farmers, making it far more accessible and cost-effective. The rapid adoption of SYMFONI on over one million hectares in the U.S. demonstrates its clear practical advantages over traditional methods.
The Vietnam Pilot Project: A Real-World Example
Vietnam offers an ideal context to pilot our sustainable practices. Rice is a big part of the economy, and it is a symbol of Vietnamese culture and the people’s staple food. Rice production contributes 30% of the country’s agricultural production value and provides food security for the nation. While grown across the country, about 50% of rice fields are concentrated in the Mekong Delta region which produces half of the nation’s rice and 90% of its exports.
With support from the World Bank between 2016 and 2022, the Viet Nam Sustainable Agriculture Transformation Project (VNSAT) trained 155,000 rice farming households on climate resilient farming and helped cultivate 180,000 hectares of low-emission rice fields. Through the project, farmers’ profits have increased by 30-35% as a result of direct savings incurred from reduced fertilizer, pesticide and water consumption, in addition to income increases from higher yields and certain price premium due to quality improvement.
As a result, the Vietnamese government has taken bold new steps in the Mekong Delta to develop one million hectares of high-quality, low-emission rice –with higher yield, better quality, and up to 70% less methane emissions.
The SIGMA MRV solution can be a critical tool in this effort, helping to accurately monitor decarbonization progress, ensure carbon credit integrity, and maximize the program’s effectiveness.
Scaling Impact Through Collaboration
A key strength of the SIGMA alliance is its collaborative model. Academia provides research, philanthropy de-risks initial investments, and the private sector is essential for scaling solutions. Partners like multilateral development organizations and governments are crucial for ensuring that these tools reach and benefit smallholders.
The SIGMA team is actively developing an ecosystem, together with Global Methane Hub, working with strategic partners to expand the uptake of SYMFONI beyond the U.S. My hope is that by working together, we can transform advanced MRV systems into actionable solutions, promoting sustainable rice cultivation and delivering scalable climate impact across Southeast Asia. It was an honor to share this vision at IDASF2025 and to highlight how the University of Illinois is at the forefront of this critical research to bring the MRV solution to Southeast Asia, starting with Vietnam as the first pilot.
Learn More
- SIGMA at University of Illinois:https://asc.illinois.edu/sigma/
- FAO IDASF 2025 Event:https://www.fao.org/digital-villages-initiative/asia-pacific/events/idasf-2025/en
- Lesly Goh’s LinkedIn Post on IDASF 2025:https://www.linkedin.com/posts/leslygoh_idasf2025-idasf2025-fao-activity-7370020731679797248-6vYk/
URBANA, Ill. — In the face of increasing global drought risk, understanding how plants respond to water stress has never been more critical. Researchers at the University of Illinois Agroecosystem Sustainability Center have delivered a pioneering synthesis that reconciles decades of differing approaches to modeling water use and dynamics in plants. The result: a unified theoretical framework and improved tools for predicting plant behavior under drought, and more informed decisions for agriculture and ecosystem management.
Transpiration, the process by which water moves from soil, through plants, and into the atmosphere, is a critical part of Earth’s water and energy cycles. It links plant physiology to weather and climate patterns and returns a significant amount of water that falls on land back to the air. During transpiration, the tiny pores in plants – stomata – release water vapor on the leaf surface and also absorb carbon dioxide to fuel photosynthesis, making transpiration a crucial indicator of plant health and productivity.
Modeling transpiration has a long and evolving history. Scientists have historically used two broad types of models to simulate plant water stress: empirical models based on statistical relationships between water in soil and plants, and mechanistic models based on physical laws that describe biological and biophysical processes. Early mechanistic models introduced empirical approaches, such as the Beta function, which scales transpiration in proportion to the soil water content. The supply-demand balance scheme is a more sophisticated empirical approach used in crop growth models that accounts for how easily water moves through soil and plant roots. These approaches are easy to implement and efficient, but often oversimplify plant physiology and lead to errors. Ideally, a model should be just complex enough to accurately simulate a specific environment.
In contrast, advanced mechanistic approaches such as plant hydraulic models offer detailed simulations of water transport through plants, but require many parameters that are difficult to measure and long computation times. Until now, the relationships among these models (and the best choice for a given application) have remained unclear.
Lead author Dr. Yi Yang and colleagues, led by the project director Dr. Kaiyu Guan, have now developed a novel mathematical framework that formally connects these different approaches. The team was able to see that the simple empirical approaches are specific cases of the complex mechanistic approach. “Scientists have debated which method is the best way to capture how plants respond to drought, but what we really needed was a clear way to connect them all,” said Yang, the former Ph.D. student of Dr. Guan in the Department of Natural Resources and Environmental Sciences. “We show that we don’t have to choose between simplicity and realism. We can understand where each method fits and see that they are connected as special cases within a unified system.”
The researchers used data from three contrasting ecosystems: a humid Midwestern cropland, a semi-arid conifer forest, and an arid grassland, to show how the best choice of model depends on the specific site and conditions. Researchers were then able to see how the models respond to different environments and different plants.
Understanding whether plant stress is caused by dry air, dry soil, or both is key. While droughts in many parts of the world are becoming more common, the Midwest U.S. often experiences ‘atmospheric drought’ (hot, dry air), even when soil moisture is still available. “This framework helps us better simulate how plants respond when the air is dry but the soil may not be,” Yang said. “That’s crucial for predicting stress under future climate conditions.”
In the example of U.S. corn and soy fields, the empirical models strongly overestimated transpiration during periods of atmospheric drought with normal soil moisture but dry air. In this case, detailed models that capture how plants manage internal water flow become more important for accurate predictions.
However, in consistently dry environments or for plants such as grasses with a fairly simple water transport pathway, the simpler Beta model works well. The arid grassland case study demonstrated that the more empirical models performed similarly to the complex mechanistic models, suggesting that the more sophisticated approach isn’t needed for reliable results in these settings.
“Our contribution allows us to determine how the drought stress affects plant transpiration with fewer data and with better efficiency,” Yang said. “If you need to rely on the more complex mechanistic model, you must gather more data and likely must go to the field to measure its hydraulic parameters in order to predict the stress level. With our improved framework and understanding, we can instead estimate that stress level with satellite data in some cases, because we can remove unnecessary parameters from the equation.”
From an agricultural management perspective, more precise estimates of plant water use help farmers optimize irrigation, saving water while protecting yield. By knowing when crops are likely to suffer stress, growers can better time irrigation during critical growth stages like mid-summer, when healthy transpiration supports peak photosynthesis and grain fill.
“How to reconcile these different transpiration modeling theories was a long-standing problem of mine from my own PhD time through my early career as a professor, ” said Guan. “I am so proud that my student Yi made such a breakthrough to close this major scientific gap.”
ASC researchers Bin Peng, Jingwen Zhang, Lisa Ainsworth, and Sheng Wang joined scientists from the University of Minnesota, Cornell University, the University of California-San Diego, University of Texas Austin and the Pacific Northwest Natural Laboratory on this project.
The project received support from an NSF CAREER award managed through the NSF Environmental Sustainability Program and USDA/NSF Cyber-Physical-System Program as well as a an annual small grant award from USGS Illinois Water Resources Center. Funding for AmeriFlux and FLUXNET data resources and core site data was provided by the U.S. Department of Energy’s Office of Science. Funded was also provided by the DOE Center for Advanced Bioenergy and Bioproducts Innovation (CABBI) funded by the U.S. Department of Energy, Office of Science, Office of Biological and Environmental Research.
The new study is titled “A Unified Framework to Reconcile Different Approaches of Modeling Transpiration Response to Water Stress: Plant Hydraulics, Supply Demand Balance, and Empirical Soil Water Stress Function” and is published in Journal of Advances for Modeling the Earth System (JAMES) [doi: 10.1029/2023MS003911].
Kaiyu Guan has been awarded the 2025 agInnovation Research Innovation Award of Excellence for his groundbreaking work in integrating advanced computational modeling with AI methods and remote sensing to advance food security and sustainability, as well as his leadership in resilient agriculture and policy-forward research.
To incentivize farmers to adopt environmentally beneficial practices, carbon credits are awarded to those who demonstrate practices that draw more carbon into the soil from the atmosphere. However, there is currently a lack of confidence that soil organic carbon credits represent real climate benefits. A research project led by Eric Potash and the Agroecosystem Sustainability Center at the University of Illinois, has shown that a more rigorous approach to soil carbon quantification is possible, one which promises to build confidence in credits representing real climate benefits.
Currently the most common approach to quantifying these soil carbon credits is called “measure-and-model.” In this approach a soil carbon project developer will measure the carbon stocks on their farms before they begin changing the practice, then they will run models on a computer to estimate the change over time.
By contrast, in the “measure-and-remeasure” approach studied by Potash and co-authors, developers measure those carbon stocks before the practice and then go back a few years later to remeasure stocks. This empirical approach can provide more reliable quantification of soil carbon accrual. Yet voluntary carbon markets and other carbon accounting approaches, including national-level greenhouse gas accounting, rely primarily on measure-and-model because of an assumption that the direct measurement approach is too expensive at landscape and regional scales.
Potash along with co-authors Mark Bradford from the Yale School of the Environment, Emily Oldfield from the Environmental Defense Fund, and ASC Director Kaiyu Guan show that measure-and-remeasure can be economically feasible for carbon crediting when a project is scaled up. The team has developed a web app, where developers can plug in a number of variables to determine how much it would cost to implement measure-and-remeasure in their projects and how profitable they can be selling carbon credits.
Instead of using biogeochemical modeling as in the measure-and-model approach, Potash and co-authors use a multilevel statistical model to estimate the costs and benefits of measure-and-remeasure. In this approach, the group estimated how much sampling would need to be done under the more rigorous measure-and-remeasure to precisely quantify the overall effect of climate-smart practices across a large number of fields. Prior academic work on soil carbon measurement hasn’t considered projects on the scale of thousands or tens of thousands of fields that occur in the voluntary carbon market.
Prior to this work, there was a perception of an inherent trade-off between rigor and economic feasibility that led most developers to take up the cheaper but less rigorous measure-and-model. In this research, Potash and co-authors have provided a framework that factors in a host of variables (all the costs and all the benefits), and shows that larger projects can be developed under the measure-and-remeasure approach and still be quite profitable. The web app enables users to interactively explore how these variables affect the economics of their specific SOC projects. Small projects can also use the app to efficiently design soil carbon measurement efforts, though they may not be profitable in the carbon market.
“Ultimately the goal is to incentivise these practices,” Potash said. “There is a huge perceived opportunity to reduce carbon emissions from agriculture and build the health of soils. At the moment, projects are being developed with measure-and-model, but we aren’t confident in their benefits. Before this research, it felt like we didn’t have another option. However, we found that there is a better way forward. Measure-and-remeasure can be economical. We think it will help to build confidence in soil carbon accounting more generally, and not just for carbon markets.”
Primary media contacts: Kaiyu Guan (kaiyug@illinois.edu), Eric Potash (epotash@illinois.edu)
Measure-and-remeasure as an economically feasible approach to crediting soil organic carbon at scale.
E Potash, M Bradford, E Oldfield, and K Guan. Environmental Research Letters. 20 (2025) 024025
Kaiyu Guan, the Founding Director of the Agroecosystem Sustainability Center (ASC) and a professor of Natural Resources and Environmental Sciences at the University of Illinois, has been selected as the American Society of Agronomy (ASA) Plenary E.T. and Vam York Distinguished Lectureship at the ASA annual meeting. The international annual meeting is the largest gathering of soil, crop, and agronomic scientists in the world.
It will be presented at the annual meeting of the ASA, the Crop Science Society of America, and the Soil Science Society of America on Nov. 10-13, 2024, in San Antonio, Texas. This conference is the most important gathering for agricultural scientists in the US and globe, for exchanging and sharing ideas, solutions, and innovation from across the field of agricultural sciences.
Guan will deliver an address titled SYMFONI: The “System-of-Systems” Solution to Quantify Soil Carbon and GHG Outcomes for the U.S. Croplands. Guan is the project leader of SYMFONI, an ARPA-E SMARTFARM project, which advanced the first-of-its-kind system-level quantification of greenhouse emissions for agroecosystems from field to continental scales. His group has developed accurate and scalable quantification of soil carbon and greenhouse gas emissions for corn, soybeans, spring and winter wheat, cotton, rice, pastureland, and miscanthus fields.
Guan founded and directs ASC, which has a mission to revolutionize agricultural systems through research, collaboration, and engagement, bridging science and practice for agricultural productivity and ecosystem sustainability. He is also the Chief Scientist for the NASA Acres Program. His research group uses computational models, satellite data, field work, and artificial intelligence to address how climate and human practices affect crop productivity, water resource availability, and ecosystem functioning. Guan’s group aims to increase our society’s resilience and adaptability to maintain sustainability of ecosystem services, food security and water resources.
The University of Illinois and the Agroecosystem Sustainability Center at Illinois are in the forefront of studying agriculture’s effect on the environment. To that end, ASC hosted “A Symposium on Agricultural Decarbonization” on Wednesday, September 18 on the Illinois campus.
The event coincided with a 1.5-day visit by Steven Singer, the Program Director at ARPA-E (The Advanced Research Projects Agency-Energy). Singer gave an overview on the program’s vision, which intersects with many of the initiatives spearheaded by ASC. Among those is the SMARTFARM program, which measures N2O and other greenhouse gas (GHG) emissions. ASC and the Institute for Sustainability, Energy, and Environment was selected as one of the SMARTFARM sites last year.
“I really enjoyed coming to Illinois and hearing about the transformative research occurring in agriculture,” said Singer. “The University of Illinois Urbana-Champaign is a leader in thinking about growing bioenergy crops sustainably, a critical aspect of developing a US bioeconomy.”
ASC Director Kaiyu Guan served as the MC for the event and spoke on “Frontier of agricultural carbon accounting technology.” He pointed to how he and his colleagues have developed a “system of systems” approach and how modeling, cross-scale sensors, and artificial intelligence are instrumental in generating accurate and scalable quantification of GHG and soil carbon change from the field to the national scales.
ASC Associate Director Wendy Yang presented insights on her research on N2O GHG emission and later moderated a panel featuring many of the speakers. John Reid, newly minted Executive Director of the Center for Digital Agriculture at Illinois, presented remarks on “Circular Bioeconomy.” Andrew Leakey, Director of Center for Advanced Bioenergy and Bioproducts Innovation (CABBI) and Chair Professor of Plant Biology and Crop Sciences, explained how the campus’ biggest research centerCABBI is leading the way in that space.
“We are thrilled to have Dr. Singer of ARPA-E visit Illinois, and we are thankful for all the support from ARPA-E,” said Guan. “Decarbonizing agricultural production to ensure both high productivity and environmental sustainability is an urgent and essential task that requires huge devotion and efforts. We at Illinois aim to lead this effort and welcome all the collaborations worldwide to join us. We expect this agricultural decarbonization symposium will recur in the coming years and aim to make this as a major event to showcase Illinois’s achievements in this space.”