Can AI help growers make smarter irrigation decisions?

Rapid technological advancements are generating mountains of data for growers to use across all aspects of crop planning, including irrigation. Yet even with all that information from soil moisture sensors, weather stations and variable-rate systems, many producers still face the same question: “When and how much should I irrigate?” 

In 2024, Nipuna Chamara, a research assistant professor at the Biological Systems Engineering Department at the University of Nebraska-Lincoln (UNL), and a team of researchers began exploring whether large language models like ChatGPT and similar AI systems can help bridge the gap between data and real-world irrigation decisions.  

The team tested the systems as part of the 2024 and 2025 UNL Testing Ag Performance Solutions (UNL-TAPS) competition. The event brings producers, seed and technology companies together to make in-season choices on seed hybrid selection, fertilizer, irrigation and chemical applications. Winners are determined based on outcomes such as profitability, input-use efficiency and grain yield.  

The role of AI in making irrigation decisions 

In the 2024 season, Chamara’s team selected ChatGPT Model 4o, fed the model key agronomic and environmental data, and used its output to guide management throughout the growing season. The UNL-TAPS program required participants to specify irrigation application rates and schedules.  

To guide irrigation decisions, Chamara’s team provided the model with: 

• weather data 

• weather forecast 

• soil moisture data 

• crop growth stage 

• irrigation costs 

• other related management activities 

• crop scouting reports 

“Considering these inputs, ChatGPT generated the date we needed to apply the irrigation and the amount, followed by logical reasoning,” he explained.  

For example, the model would suggest applying one inch on Tuesday or waiting until the following Thursday, based on the uploaded data. 

“With the low commodity prices and high energy prices, ChatGPT was helping with decisions around whether it is really worth irrigating another inch or not,” Chamara said. “That is something that a soil moisture sensor cannot tell you.” 

Results from the UNLTAPS competitions 

In 2024, the AI-managed farm finished strongly, recording the seventh highest yield out of 30 participants, according to Chamara. However, the irrigation efficiency was below average, and the AI-guided team required more water to produce its yield outcomes. 

In 2025, the team adjusted its AI use based on the previous year’s outcomes. They tested their updated approach across four UNLTAPS competitions in North Platte and Mead, Nebraska, including sprinklerirrigated corn and sprinklerirrigated soybeans. 

“In the sprinkler-irrigated corn contest we had the highest yield out of 30 participants,” Chamara said. “In two other sprinklerirrigated corn competitions, we finished in the top 30%, and in the soybean competition, we were in the top 50% for yield.” 

The researchers attribute part of that improvement to better models and better inputs. 

Practical tips for prompting and privacy 

For growers interested in trying these tools, Chamara offers these three tips: 

1. Use strong prompts. Prompt engineering, or how you ask the AI to respond, is essential. Follow-up questions also improve results. Growers can ask the system to explain its recommendation, compare alternate irrigation scenarios or adjust its response based on updated weather or soil moisture data. 

1. Use the model’s multimodal capabilities. AI tools can process and understand many different data types, including text, images, audio, video and even weather forecasts, soil health reports and soil moisture readings. 

1. Select tools carefully. Free versions are available, but Chamara suggests considering the paid versions. 

“The free versions can run into usage limits during time-sensitive decisions, and you will have better data privacy with paid versions,” he said. 

After two seasons in UNLTAPS competitions, the researchers believe AI tools can help growers interpret data faster, ask better questions and make more informed decisions about when and how much to irrigate.