Precision Agriculture Tech Can Address New Fertilizer Shortages

The ongoing instability in the Middle East has put fertilizer availability at the center of global food security concerns. Up to 30 percent of the global fertilizer trade typically passes through the Strait of Hormuz, along with major flows of liquefied natural gas, a key feedstock for its production. Delayed shipments are rippling through agricultural supply chains, and the United Nation’s Food and Agriculture Organization has warned that the price of urea, a widely used nitrogen fertilizer, had increased by 52 percent in the United States and by 60 percent in Brazil by mid-April.

“Traditionally, because fertilizers are relatively cheap, farmers often apply the maximum allowed amount,” says Ioannis Athanasiadis, a professor at Wageningen University in the Netherlands who works on AI for agriculture and food systems. “It acts as insurance against uncertain weather conditions.” But for farmers already squeezed by the cost of fuel, machinery, and seeds, volatile fertilizer prices are making waste increasingly expensive.

Precision agriculture has already helped farmers reduce chemical use and save money, with computer-vision systems able to identify weeds and trigger herbicide nozzles only where needed. But fertilizer is a tougher challenge. Nitrogen, the key nutrient in many fertilizers, is invisible, highly mobile in soil, and can be washed below the root zone before crops absorb it.

“The difficulty is that you never actually know how much nitrogen the plant and the soil have,” says Chris Padwick, a technical fellow at Blue River Technology, a California-based company that develops computer-vision technologies for agriculture.

Precision Fertilizer Technologies

Along with its parent company John Deere, Blue River developed a precision fertilizer technology called ExactShot, which can be deployed only at the time of planting. The system detects each seed as it goes into the soil and sprays a few drops of starter fertilizer directly onto it, instead of applying fertilizer continuously along the row. Blue River says the system can cut starter fertilizer use by more than 60 percent and could save more than 93 million gallons annually across the U.S. corn crop.

The harder task comes later in the plant’s life, when crop needs depend on weather, soil type, previous applications, and what has already happened in that patch of field. Many sprayers for precision herbicide applications can be fitted with attachments in the shape of an inverted-Y that drag hoses near corn plants, dribbling liquid nitrogen close to the row during the growing season. But without reliable information on which parts of the field actually need nitrogen, the application remains nonselective.

John Deere’s ExactShot technology detects each seed as it is planted and applies a few drops of starter fertilizer.Blue River Technology

Padwick says Blue River is testing whether its crop imaging systems could also become “digital scanners” for plant health, using broader spectral coverage to analyze vegetation as sprayers pass close to the canopy. But getting insights isn’t the same as accurately interpreting them. Yellowing leaves or reduced chlorophyll may point to a nutrient deficiency, but they can also indicate drought stress, disease, or insect damage. That is why some companies are moving below the canopy and into the soil, where direct measurements can provide a more objective picture of what nutrients are available.

Elsewhere in the United States, Iowa-based N-Sense is betting on a mobile machine for soil analysis. Its prototype soil-nitrate sensor can be pulled through the field by a truck to measure nitrate concentration in real time. The system uses a ruggedized miniature Fourier-transform infrared spectrometer operating in the mid-infrared, coupled with a diamond interface. Soil is pressed against one end of the diamond while infrared light passes through the other, allowing the instrument to detect nitrate while the diamond protects the optical surface from abrasion.

“Nitrate is particularly difficult to detect,” says David Laird, N-Sense’s president and CEO. “In the ultraviolet, many things in the soil are responsive, so it is very difficult to separate the nitrate signal. But in the mid-infrared, we are able to isolate the nitrate band and get a strong signal.”

N-Sense’s sensor measures nitrate in real time as it moves through a crop field.N-Sense

The sensor feeds nitrate data into machine learning software that also taps into soil-survey data, satellite imagery, and yield data to generate fertilizer prescriptions that can be uploaded directly to a tractor.

“You do not want to look at nitrogen in isolation,” says Laird. “The soil may have low nitrogen content, but if water availability is what is limiting productivity, adding nitrogen will not help.”

In one field tested last year, Laird says the company achieved about a 30 percent reduction in total nitrogen fertilizer applied.

Real-Time Soil Analysis

In Potsdam, Germany, an agricultural tech company called Stenon has developed FarmLab, a mobile soil-analysis device designed to give farmers real-time measurements directly in the field. The handheld probe is pushed into the soil and combines optical spectroscopy, which reads how soil absorbs and reflects light, with impedance-based electrical measurements, which send a small electrical signal through the ground to capture properties affected by moisture, salts, and nutrient ions. Environmental sensors capture the temperature and humidity, and there is also a GPS tag that associates each reading with a location. Cloud computing and machine learning turn the raw signals into usable soil data. The goal is to infer key soil parameters, including nitrate, mineral nitrogen, moisture, and other indicators that can guide fertilizer decisions.

Stenon’s FarmLab uses optical spectroscopy, electrical impedance sensing, and machine learning to generate soil nutrient maps for precision fertilization.Stenon

The company’s founder and CEO, Niels Grabbert, says the idea came when Europe began enforcing the Nitrates Directive, a law aimed at protecting groundwater and surface water from agricultural nitrate pollution. Back then, farmers were required to reduce nitrogen losses but lacked real-time information on the nitrogen present in their fields.

“Depending on the country and on what parameters you are testing for, it can take anywhere from two to eight weeks before you receive soil-testing results,” says Grabbert. “That means farmers do not know in real time how much nutrition the soil itself can provide.”

FarmLab is meant to replace sparse lab testing with faster, denser field data. On a 100-hectare farm, an agronomist could take one reading every two hectares, then use the software to turn those GPS-tagged measurements into nutrient maps and fertilizer rates that can be sent to farm machinery or management platforms.

Grabbert says the technology can reduce fertilizer use by around 20 percent on average while increasing yields by 2 to 8 percent, depending on the crop and production system.

For Dutch professor Athanasiadis, these systems point in the right direction, but they are not enough on their own. “There are no magic solutions,” he says. “We need sensors, robotics, AI, government support, and farmer participation all working together.”