Skip to main content

Artificial Intelligence for Irrigation Models

Mohammad Valipour’s vision for artificial intelligence irrigation systems spans all of the world’s farms — from the small family-owned land tracts — to vast commercial grove and field operations. His ambition sprang up as a youth in western Iran. He lived about 600 kilometers from Tehran, in the valleys near Kermanshah, in a rural town called Eslamabad-e Gharb. There, families speak Kurdish and make a living from crops they produce on modest farms. Valipour’s family is a member of the Kalhor tribe in Iran, one of the most ancient and powerful, according to scholars who contribute to the Encyclopaedia Iranica.

In Eslamabad-e Gharb, the climate is mild, semi-arid, and the valleys are green most of the time. The soils are slate and mocha-colored. Rock shelves form the toes of mountains. On the face of one rock shelf is Taq-e Bostan, ancient arches that protect bas reliefs of Persian Sassanid art from the fourth century. Higher than the repeated rock shelves is Mount Paraw, which holds some of the world’s most upper and deepest caves.

Valipour recounts memories when, along with his family, he would scoop green chickpeas from their pods and load them into his mouth. “Tasting the freshness of sweet chickpeas is so different from eating them from a can or a frozen box,” he said. “When you taste something freshly harvested, it is intense.”

Valipour’s family grew chickpeas and wheat in alternating years. Today, his cousin continues with the farming tradition. The land inherited from previous generations has been in the family for longer than anyone can remember. There are no records. However, Valipour has memories, and not all of them are positive. At times, there was no water in the soil or pests emerged, and the family’s crops would perish.

“We had not had enough yield, productivity was low because we had no access to water to irrigate the field,” said Valipour. “Our fields were watered only by rain. Also, sometimes there was only minor rain. Other times, there were droughts. When there was no water, we knew we could not expect enough yield.”



The moments during which Valipour’s family lost their crops changed him. He wondered how he could extract water from the land. It was those moments that moved him to his career – a lifelong commitment to investigate irrigation sustainability for more reliable crop yields. Over time, and through his formal education, Valipour learned the science behind hydrology and irrigation. He said he would use his skills to help his family, his neighbors — and his nation.

“Water affects everything, and for crops, water is especially important for yield and productivity. I decided to study irrigation and water science to gain knowledge about water management processes for crop production,” said Valipour.

Today, he is a highly trained hydrologist and irrigation scientist. His vision to improve crop yield with smart irrigation continues with his research and findings. The findings must be transferred as clearly as possible to growers, he said.

“Irrigation operations must be adapted to consider climate change and water shortages to achieve sustainable crop development in the future,” said Valipour.

Valipour earned a bachelor’s in irrigation from Razi University, a master’s in irrigation and drainage from the University of Tehran, and a Ph.D. in Water Sciences and Engineering from Sari Agricultural Sciences and Natural Resources University. All of the forgoing universities are in Iran.


For his undergraduate degree, Valipour’s work involved quantitative analysis of land topography using computer programming. During his master’s and doctorate studies, Valipour’s work took place indoors, with models and software. He learned to make simulations – or to predict possible outcomes for hydrology scenarios using computers and historical data stored for past decades. He used models to simulate irrigation and hydrology. Models help us to identify changes in crop water requirements. Water needs change based on variations on weather variations such as heavy rain in Florida, hurricanes, or a dry summer in Iran.

Valipour recalls a statement made by renowned British statistician, George E. P. Box: “All models are wrong — but some are useful.”

“What this means is that although modeling can help us to have an insight of natural phenomena, nothing can predict with 100% certainty what is going to happen in the future,” said Valipour.

During his doctoral studies, Valipour used artificial intelligence, or more specifically, genetic algorithm and gene expression programming. The pattern evolves and automatically creates computer programs that tell irrigation equipment to dispense certain water volumes. The programs then mimic natural responses like a tree taking up water or losing moisture in its leaves during a hot, sunny day.

Valipour developed hybrid models to estimate evapotranspiration over agricultural fields in Iran. He included climate change projections to determine the best conditions for crop production. Valipour said natural-based models, “neural networks” are inspired by nature. “Traditional mathematical models loop redundantly when you want to solve complex non-linear problems like plant responses to water,” said Valipour.

Valipour codes decades of historical data about weather and water availability. He then uses data to predict future needs when specific weather conditions emerge. For his master’s work, he forecast the amount of water that would flow into a dam using artificial neural networks.

When he graduated with his Ph.D., he decided to extend his knowledge in the U.S. and learn about humid climates and technologies for crop production.


As a post-doctorate at Auburn University in Alabama, Valipour learned how to analyze data from satellite images. The images provide historical data for weather and hydrology to his models. In America’s deep balmy south, growers have vastly different irrigation needs than do Iran’s farmers. Alabama soils are fertile and retain more water.

During Valipour’s studies in Alabama, he evaluated soil moisture dynamics using remote sensing, land surface models, climate reanalysis and in situ measurements, or field data. He also assessed the interactions between soil moisture and precipitation and evapotranspiration over the entire U.S. by using soil moisture and root zone soil moisture estimation technology produced by NASA, called Soil Moisture Active Passive, or SMAP.

“The results showed that one of the remote sensing methods, SMAP, had better agreement with soil moisture observations collected by sensors in the field. Furthermore, most of the soil moisture products show significant correlation with precipitation and evapotranspiration.”

The computer and software-generated models gave him the predictability he wanted for his career, but he wanted to go further and understand how data is created by working on field-based experiments. He wants to produce meaningful predictions based on the relationships that you can only see by walking through the field. Searching online, he found Dr. Sandra Guzmán’s smart irrigation and hydrology lab at the University of Florida Institute of Food and Agricultural Science’s Indian River Research and Education Center (UF/IFAS IRREC) in Fort Pierce.

“My background is in computer simulation, but I also want more knowledge in plants and field conditions,” he said. “There is a considerable difference between what you think you can simulate with models and the dynamic interrelations you find in the field.”

Valipour said the main irrigation problem for both Iranian and American farmers is that surface irrigation is inefficient when it is not managed correctly.

Growers are not sure how much or for how long to irrigate. Even with the best climate, important conditions are unknown, especially the way the soil and plant interact with water. One of those examples is irrigation management when citrus greening is present. Also, the type of soil and irrigation system is important to provide high efficiency in water use.

“The important thing is that we can improve,” said Valipour. “Now we have pressurized irrigation and smart technologies such as controllers and soil moisture sensors to operate efficiently, but still a lot is needed.”

For this reason, improving irrigation efficiency is essential, so that farms may operate with higher efficiency. A properly installed and well-managed set of sensors and controllers could provide huge savings in water and energy use, said Valipour.


Valipour’s work as a post-doctorate at IRREC entails rapid development of new irrigation technology for Florida’s citrus groves. His work in Florida is part of his vision to modify and improve irrigation. A second focus for his work is evapotranspiration.


“Here at IRREC, I have focused on two different projects. The first is the evaluation of best management practices and smart irrigation to make more efficient the use of water and nutrients for citrus production. I am evaluating citrus beds covered with a fabric mulch and making daily irrigation recommendations based on an automatized irrigation system. My second research focus is to assess variations in evapotranspiration based on a set of related climatic variables,” said Valipour.

Valipour said he can develop quick recommendations for citrus growers to irrigate young trees growing under groundcovers. The guidelines will be specific to each grove’s conditions and are analyzed based on data gathered from soil moisture sensors installed in the field and information provided by the grower.

“Hydrologists are able to help food producers everywhere in the world with the development of more efficient irrigation practices,” said Valipour. “There is much work to be done, and we will keep working to achieve greater food production while conserving water.”

One Comment on “Artificial Intelligence for Irrigation Models