“When people think of food, they think it comes from Publix, or McDonald’s,“ Gerrit Hoogenboom, Ph.D. muses, “but in order for the food to get there, it needs to be grown or produced by farmers”
Most people think of agriculture as just agriculture, but producing food for a dramatically increasing population is a complex and sophisticated endeavor. On a rapidly changing planet, the demands for certain crops are harder to keep up with. If the weather changes by even a few degrees, farms that could once consistently produce regional crops may find that those crops are no longer viable or cannot thrive. A difference in the amount of rainfall, the amount of heat can change how plants grow, and those differences also can create environments for different pests and diseases to thrive. In a changing global landscape as such, where every part of the equation is a variable, developing crop modeling software that can support farmers with real time feedback will be a critical tool in providing food to, well, all people. In a feverishly shifting climate, the urgency to deliver agricultural advice quickly is unmistakably high.
Hoogenboom, a Professor and Preeminent Scholar for the Agricultural and Biological Engineering (ABE) Department, specializes in the development and application of crop simulation models and decision support systems. He coordinates the Decision Support System for Agrotechnology Transfer (DSSAT), one of the most popular crop modeling systems across the world, an elaborate web of data that provides agricultural information for users in over 190 countries.
It all started in the 1980s when the U.S. Agency for International Development (USAID) invested $10M in a computer model that would help farmers in low-income countries optimize their crop choices.
Those early crop models weighed individual variables like rainfall, temperature, and yield. The current, complex DSSAT model expands on simpler versions to include a more robust overview of agricultural variables, “We are looking at agriculture as an entire system. Looking not only at the crop but also looking at the weather and soil, in addition to inputs, such as fertilizer, irrigation, and crop varieties that farmers can adjust. This is a systems-analysis approach that incorporates software, data, and human knowledge. The model integrates weather inputs, soil data, crop genetics, and management practices and delivers predictions about yield, harvest time, and potential interventions,” he explains. Hoogenboom is also collaborating with another Ph.D. in the ABE department, Willington Pavan, Ph.D., to consider the impacts of pests and disease, adding those data to the DSSAT model.
The software is complex and high level, so the information is best utilized by extension agents, crop consultants, researchers, and educators, “The software itself is not very user friendly because we haven’t had the financial resources to bring it up to speed to the current standards. But indirectly, the outputs of the models can be used by growers and producers, or NGOs or private companies,” Hoogenboom explains.
The goal is eventually to integrate the software into an interface a farmer standing in a field could refer to, to make a real time decision, and there are already emergent programs that do just that.
The DSSAT project requires collaborative intelligence sourcing and a hive mind approach. Annually groups of engineers, agronomists, soil scientists, and others meet in organized week-long hackathons, collaborating often on a volunteer basis to plug in up-to-date data to keep the model accurate. “In the background we need to basically obtain all the input data to run the model, to find local weather data and soil information for that location based on the latitude and longitude,; we need to know a little bit about the farmer’s management, whether and how they are irrigating and fertilizing for example, and we also need to know what type of cultivar or hybrid they are growing there,” Hoogenboom said.
The future of global food security depends on predictable, healthy crop yields, and the key to that aim is better risk management, education, and the ability to disseminate that information broadly via accurate and readily usable information.
4
