The future of agriculture

How South Dakota State University researchers are using cutting-edge technologies to move the agricultural industry forward. 

Editor's note: This article was featured in SDSU's research magazine, STATE of Discovery. 

It's a mild early spring morning at the historic Cottonwood Field Station in western South Dakota, and a herd of 150 Angus steers are scheduled to move to a new pasture rotation. Moving cattle can be tricky and often requires some extra help, electrical fencing and quite a bit of time. But today, there are no extra ranchers, no gates swinging open and no temporary fences in place.

Instead, South Dakota State University assistant professor Hector Menendez is sitting comfortably in his office, miles away from the pasture, looking at a satellite image of Cottonwood's 2,500 or so acres of rangeland. While sipping on his morning coffee, he draws the boundaries for a new pasture on his phone. By the end of the day, the herd of cattle will have arrived at their new location.

"It's a pretty easy application to use," Menendez said. "The technology will guide the animals to their new pasture."

This is the virtual fencing technology that has become the "hot topic" at livestock management conferences, meetings and workshops the last few years. Producers, faced with rising input costs, increasing labor needs and mounting environmental pressures, are eager to learn about an emerging technology that has the potential to address some of the industry’s biggest challenges.

At ranches and fields across South Dakota, SDSU researchers are actively developing and investigating how virtual fencing and a wave of other technologies, including remote weighing, artificial intelligence and drones, can support the state's producers and reshape modern agriculture.

A fenceless landscape
 
At SDSU, researchers have been conducting studies on the use and utility of virtual fencing since 2020. This technology is being increasingly adopted by ranchers across the country, and some believe it has the potential to revolutionize beef production. But how does it actually work?

The technology is similar in nature to pet fencing for dogs. But instead of an underground wire to contain the animals, geofencing technology via satellites and wireless networks is used. Each of the steers in the herd have been fitted with a GPS collar. When Menendez draws a new boundary, the collars communicate with a solar-powered base station on the property, which defines the boundaries of the pasture. When an animal approaches the virtual boundary, the collar will emit a pinging noise. If the steer continues, it will receive an electric stimulation. If the animal passes through the virtual boundary, the system will allow it to return to the pasture without additional stimulus.

SDSU researchers first became interested in virtual fencing technology when looking at ways to better implement rotational grazing strategies. Various research studies have found that rotational grazing strategies increase the available forage, allow for higher stocking rates on the landscape and improve soil health. All of these factors provide benefits to the rancher and increases the economic potential of their production system. The problem, however, is that implementing rotational grazing strategies requires physical fencing to separate pastures on a ranch. In South Dakota, one mile of fencing can cost well over $15,000. On top of the costs, ranchers need to physically move their herds from one pasture to another, which is time-consuming and requires a significant amount of manpower.

"Input costs are the main drivers for producers when making decision on their production systems," said Elias R. V. Moreno, a doctoral candidate working under Menendez's guidance. "That's why we are looking at virtual fencing, as well as other precision livestock technologies, as a new way to implement rotational grazing strategies."

Virtual fencing, first conceived of in the 1990s, is an ideal technology for implementing rotational grazing strategies because it removes a lot of the manpower requirements for moving herds from one pasture to another and cuts down on input costs, like fencing.

At Cottonwood, the SDSU research team has completed comprehensive studies investigating the effectiveness and pitfalls of virtual fencing on beef production systems. According to the team's findings, virtual fencing does not negatively impact livestock production.

"There were no differences in average daily gain, dry matter intake, feed conversion, marbling and carcass quality, when comparing continuous grazing to virtual fencing enabled rotational grazing," Moreno said.

The technology provides benefits beyond beef production. Virtual fencing gives producers "peace of mind" and allows them to leave the ranch while still being able to make management decisions. In the Black Hills, virtual fencing has been a boon for producers on large-scale ranches with rolling topography. The GPS collars allow them to track exactly where their cattle are, significantly cutting down the amount of time it takes to find and round-up a herd on a 20,000-acre ranch.

"Virtual fencing gives ranchers a reassurance of where their animals are at all times," said Ira Parsons, assistant professor of animal science at SDSU. "The technology allows ranchers to check in on their animals, even if they are not physically on the ranch. It really allows for new opportunities and even a different lifestyle as a rancher."

As Jameson Brennan, SDSU associate professor and SDSU Extension livestock grazing specialist, notes, virtual fencing can help improve the work-life balance for ranchers. It can also potentially help get people back into the ranching business.

"I can be at my kid's ball game and pull out my phone to check on where my cattle are," Brennan explained. "Or for people with a full-time job who want to get back into ranching but want to reduce the labor needs, virtual fencing provides a good pathway back into the business."

While most types of perimeter fencing are required by law in South Dakota, virtual fencing may reduce or even eliminate the need for some interior fencing needs on ranches, Parsons noted. Researchers believe this may help wildlife populations, like antelope and mule deer, whose movement and migration corridors can be restricted by traditional physical fencing.

Virtual fencing also has the potential to improve water quality. In the western U.S., livestock and wildlife depend on prairie streams and reservoirs as their primary water sources. Riparian vegetation — the plants near riverbanks — are extensively used by livestock and provide a disproportionate amount of available forage on a pasture. In many places, riparian areas have been overgrazed and damaged, leading to water quality issues in prairie streams. Virtual fencing technology can help make riparian areas "exclusion zones" for cattle, allowing the vegetation time to recover. According to Brennan, who conducted a study on virtual fencing and riparian restoration at Cottonwood, the technology is "99.5% effective" at keeping them out of exclusion areas.

"It works really well at keeping them out of the creeks," Brennan said.

In South Dakota, the number of virtual fencing adopters is increasing, and Brennan knows a handful of ranchers who have fully implemented a virtual fencing system on their property. SDSU researchers and Extension specialists are continually holding training sessions and workshops to help producers decide if virtual fencing is right for their system.

"The fundamental question producers have to ask themselves is what is their ultimate objective?" Brennan said. "There's always going to be some folks who maybe don't need virtual fencing for their management goals. But for a lot of other folks, virtual fencing provides a solution that can really bring some benefits to the ranch."

Remote weighing 

Virtual fencing is not the only precision livestock technology SDSU researchers are investigating. At Cottonwood, Parsons is utilizing remote weight capture systems to collect large amounts of data on his herd. The data than can be used to inform management decisions on the ranch.

"It is a very useful research tool," Parsons said. "What we use it for is to really monitor and fine-scale animal changes in weight."

Body weight is a critical component for monitoring animal body mass, body condition, nutritional status and health. However, traditional methods of collecting body weight are stressful, costly and logistically impractical with cattle on extensive landscapes. Parsons is utilizing a walk-over-weigh with a platform scale system, which wirelessly collect weights from cattle out in the pasture. These weights can help develop and inform nutrition models for rangeland cattle.

The value in all this collected data — whether from daily weight measurements, GPS tracking via virtual fencing or climate information from the Mesonet — Parsons says, is when it is integrated with each other. Machine learning, a type of artificial intelligence, can fuse all of this data together to gather meaningful insights for producers and researchers.

"American agriculture is an extremely efficient system," Parsons said. "We are at the top of the world in doing what we do. By putting more data in the hands of producers, we can move the needle and become even more efficient and productive."

With these precision livestock technologies, Parsons believes that producers can make more fine-scale management decisions with data-driven tools, improving their productivity and efficiency while reducing their inputs costs and labor requirements.

"These tools can help producers reinforce what they are seeing with their eyes and what they need to change in order to improve the performance of their cows," Parsons said. "The idea is not to replace producers and institutional knowledge; these tools add affirmation to what they are seeing and give them clues on what they need to keep an eye on."

Aerial coverage

Weeds have plagued farms for as long as agriculture has existed. In South Dakota, three dominant types of weeds — waterhemp, kochia, and foxtail — compete with cash crops, like corn and soybeans, for sunlight, nitrogen and water. When weeds arrive early or in the middle of the growing season, they can cause significant yield losses. Uncontrolled weeds can cause up to 50% yield loss in corn fields. This can equate to millions in total losses for South Dakota's producers.

To control weeds, producers will often apply herbicides using tractors with large sprayers. While this can be effective in uniformly applying herbicide across a large field, the wheels of tractors can compact the soil, reducing the pore space in the soil needed for water, oxygen and root development. Research has found that soil compaction can reduce yields by as much as 20% under certain conditions.

Researchers in SDSU's Department of Agricultural and Biosystems Engineering are developing creative solutions to weed management by taking herbicide application to the sky.

Drones, or unmanned aerial vehicles, have become increasingly present in agricultural research and production systems. Pappu Yadav, assistant professor of precision agriculture, and his team are building a custom-fitted “sense-and-act” hybrid drone. Unlike off-the-shelf models that spray uniformly, this drone uses an edge-AI powered computing system along with RGB (red, green and blue) and multispectral cameras linked to a precision single-nozzle sprayer. This allows the drone to identify a weed and neutralize it in a single pass, a sort of "search and destroy" mission for unwanted plants.

"By identifying weedy spots, farmers can apply targeted interventions to optimize the use of agricultural resources," Yadav said. "This enables them to make proactive decisions that minimize waste, enhance sustainability, prevent further damage and maximize yield potential."

The research team, which includes undergraduates, graduate students and research associates, is working collaboratively to build out the drone. The computing system — the "brain" of the operation — is being developed by research associate Krisha Joshi, who utilizes edge-AI to bridge the gap between raw data and agricultural action. By training AI models on massive datasets of South Dakota weed species, the system can distinguish between a valuable corn or soybean plant and a thirsty waterhemp instantaneously. 

Drones equipped with spraying devices are already on the market. The model the team is using is from Hylio AgDrones, a Texas-based company, has four nozzles for herbicide application in its standard buildout. Devraj Majhi and Dustin Seubert, two undergraduate researchers in the lab, customized this drone with a single nozzle — for targeted precision spraying — and used a 3D printer to create the components needed to secure the edge-AI computing device, cameras and other attachments to the drone.

"The angle of the nozzle and the drone's flight altitude are critical," Majhi said. "We have to account for the downward air force from the rotors, to ensure the herbicide hits the weed and doesn't drift onto the valuable corn or soybean crops.

Precision application will help producers cut down on input costs and may also provide some environmental benefits. When herbicide is overapplied to fields, it can run off the soil and into water sources, like rivers and lakes. There, herbicides, alongside synthetic fertilizers, can contribute to toxic algal blooms, which turn lakes green. Research has shown that herbicides can also contribute to soil degradation as the chemicals can kill bacteria and fungi needed for nutrient cycling.

"Precision spraying can help farmers cut down on input costs while improving soil health," Seubert said.

The drone is designed to be user-friendly, despite the high-tech look.

"It can fly fully automated," said graduate student Aastha Gautam, who is working on her master’s thesis using the hybrid drone. "From takeoff to landing, the 'flying' aspect is less complex than it looks. Users simply plan the drone’s 'mission' — its flight path — before it ever leaves the ground."

While the drone is currently optimized for targeted precision spraying, its true versatility lies in its modular imaging payload. SDSU’s Machine Vision and Optical Sensor team, led by Yadav, is developing a system that allows sensors to be swapped out based on the specific needs of the season. Currently, the drone utilizes a combination of RGB and multispectral cameras to map weed infestations and monitor general crop health. However, the modular design means that as research evolves, the sensors can be easily replaced with thermal or hyperspectral cameras for specialized tasks like irrigation scheduling or high-fidelity crop disease and pest detection.

"Our research focuses on developing the underlying algorithms so that the hardware can remain flexible," Yadav explained. "Whether it’s identifying subtle spectral signatures of a specific disease or using thermal data to detect water stress, the goal is to provide a platform that can adapt to any agricultural challenge."

By flying these missions, the drone generates a high-resolution heatmap of the entire field. The multispectral sensors capture wavelengths of light, including near-infrared, that are invisible to the human eye but highly indicative of plant vitality. When combined with the team's machine learning algorithms, this technique allows producers to catch trouble in their fields days or even weeks before it would be visible from the tractor.

"Our data analysis techniques and AI-driven algorithms transform raw data into actionable insights," Yadav said. "By creating a modular, automated system, we are ensuring that South Dakota producers have access to the most advanced tools available, both now and in the future."

Ag transformation

Since the birth of agriculture, scientific and technological advancements have propelled the industry forward, enabling producers to become more efficient and productive as time goes on. Early hand tools transformed food production and allowed societies to form. In the 19th century, Joseph Glidden’s barbed wire invention revolutionized ranching across the American West. The mechanization boom of the early 20th century reshaped farm labor and productivity, and the Green Revolution introduced synthetic fertilizers and high-yield crop varieties that dramatically expanded the global food supply.

Today, another transformation is underway, one built not by steel and wire but by satellites, sensors and AI.

From virtual fencing and remote weight monitoring systems to hybrid drones and multispectral cameras, SDSU researchers are developing and evaluating technologies designed to help producers save time, reduce labor demands and make more precise, data-informed decisions. These tools are not replacing producers’ experience, instinct or knowledge. Instead, they are strengthening decision-making through accurate data points and providing new layers of insight that allow for more responsive, efficient management.

Agriculture has always evolved alongside technology. The pivotal work being done by SDSU researchers ensures South Dakota remains at the forefront of the industry's current evolution.