Emerging Precision Agriculture Tool Improves Grazing Systems  

This year, UCCE Central Sierra has partnered with the Bureau of Land Management (BLM) and the Foundation for America’s Public Lands to bring virtual fence to BLM lands here in the Sierra Nevada. As part of the Lands to Love Catalyst Fund grants program, UCCE researchers received one of two grants awarded in California to kickstart this exciting new project. 

The project will take place on Hunter Valley Mountain, a BLM allotment near Lake McClure in Mariposa County.  VF technology will allow livestock to be contained exactly where they are needed on the landscape with greater flexibility, deployment, and intention than traditional fencing allows.  Precision grazing through virtual fencing will allow for more accurate livestock location monitoring, exclusion of livestock from sensitive habitats, targeted grazing to reduce fuels or invasive weeds, and minimizing conflict with recreation. One of the overarching goals of this project is to demonstrate how the flexible and adaptive qualities of virtual fencing can strengthen partnerships between livestock producers and federal land managers, and help make sustainable grazing and land stewardship goals on multi-use public landscapes more attainable. 

Strategically placed buffers help slow the spread of wildfire 

California’s annual rangelands are famous for their green, rolling hills in winter and spring. But as the grass dries out, these landscapes become highly flammable and pose a significant fire risk. According to CAL FIRE’s Wildfire Activity Statistics Annual Reports, grass fires were the most common type of vegetation burned in the Central Sierra counties of El Dorado, Amador, Calaveras, and Tuolumne. From 2019 to 2023, they accounted for approximately 78% of all acres burned in the region, averaging about 3,033 acres annually (CAL FIRE, 2019–2023).

Mitigate wildfire risk by reducing fine fuels and adding fuel breaks

Reducing fine fuels is a critical proactive step in mitigating wildfire risk, especially in high-priority areas where grasslands border human activity. Fuel breaks are often most effective when placed near likely ignition sources, such as roads or powerlines, and around vulnerable assets like homes, communities, or critical infrastructure. In both cases, these strategically placed buffers help slow the spread of wildfire, improving the chances of containment and reducing the risk of catastrophic damage. Every year, fuel breaks are created and maintained by private landowners and livestock producers, county road and public works crews, local Fire Safe Councils, Resource Conservation Districts, CAL FIRE, private contractors, and others. 

Methods for creating and maintaining effective fuel breaks include targeted grazing

Common techniques include mowing, herbicide application, prescribed fire, grading, cultivating, and targeted grazing. Choosing the right method depends on many factors, including site conditions, ownership, ecological concerns, and available resources.

Targeted grazing can make a lot of sense on annual rangelands because it leverages livestock’s natural foraging behavior to turn fine fuels into marketable weight. This reduces labor for landowners and generates income for producers. Fuel breaks can protect ranchers from losing feed to fires that start along roads, and just as importantly, they help protect surrounding communities from fires that may start on the pasture. However, implementing fuel break grazing typically requires fixed or temporary infrastructure, often electric fences, parallel to the outer perimeter fence to concentrate grazing for the desired fuel reduction.

Virtual fencing as a flexible, labor-saving alternative to fenced fuel breaks

In June 2024, UCCE tested whether virtual fencing (VF) could provide a more flexible, labor-saving alternative to fenced fuel breaks by eliminating the need for an inner physical fence while still achieving comparable fuel reduction. The trial was conducted on a privately owned pasture adjacent to the city of Sutter Creek. After a productive growing season, the grass stood over 3 feet tall in many areas and averaged 4,269 pounds of dry forage per acre (the equivalent to about 71 bales of hay per acre, assuming 60 pounds per bale). Several homes stood within 50 feet of the fence line, making the fuel load a clear fire hazard. Removing such fuels could help achieve the 100’ of defensible space that CAL FIRE recommends.

Thirty-seven cattle of mixed age and breed with no prior exposure to VF were selected for this trial (Figure 1 Left). A single VF base station was placed near the trial site to allow the collars to properly communicate. After an 8 day training period, the herd was confined to a 150 foot wide grazing area between the outer hardwire perimeter fence and the VF boundary. The herd received salt and protein supplement during the trial to account for the decreasing forage quality of the dry grass. These were placed away from water and loafing spots to encourage more uniform utilization. Over 19 days, the herd grazed the 7.7 acre fuel break down to 780 pounds per acre (about 13 bales of hay per acre), an 82% reduction of flammable fuels (Figure 1 Right).

Figure 1. Left: Herd wearing VF collars. Right: Results of the fuel break grazing. The dashed line shows where grazing stopped along the VF boundary. 

The herd respected the VF boundary 99% of the time (Figure 2), being contained by the audio cues alone 81% of the time (Figure 3). Visually, the livestock appeared calm throughout the trial. The only water in the pasture were troughs placed inside the fuel break. As VF boundaries only prevent animals from leaving an area, the few cattle who did escape would eventually return voluntarily, drawn either by thirst, supplement, or the presence of the herd.

Figure 2. GPS locations (blue dots) of the herd during the 19-day trial. These points outline the location of the fuel break in relation to the city of Sutter Creek. 

Figure 3. Top: Percentage of audio and electric pulse cues received by the herd each day. The livestock learned to respond primarily to the audio cues alone as the trial proceeded. Bottom: Audio and electrical pulse count for the 37 member herd per day. Fuel Break 1 was the smallest and therefore required the most cues for containment. Given the herd size of 37 cows, ~600 cues per day equates to about 1 cue per animal every 1.5 hours.

Virtual Fencing offers extraordinary levels of flexibility compared to traditional fencing

Compared to physical fencing, VF offers extraordinary levels of flexibility to adapt grazing and exclusion areas to meet ongoing management needs. In this trial, a newly paved road crossed the fuel break, so we split the fuel break into two VF zones to exclude livestock from the road. The herd was easily moved to the second fuel break after grazing the first.

The primary tradeoff with using VF in this trial instead of electric fence was the width of the fuel break. While John Allen, the participating rancher in this trial, typically installs 60 foot wide fuel breaks using electric fencing, we chose a 150 foot wide VF to account for GPS inaccuracy and to reduce stress on the animals from excess audio and electric pulse cues. While this wider area provided a greater level of protection from wildfire, it also took approximately twice as long to graze to the same level of fuel reduction. 

“VF is quicker, easier, and more reliable than the electric fence we use nearby to graze a firebreak along the roadway. Also, it’s probably better suited to cows rather than calves that are going to the market,” says rancher John Allen.

Trial suggests Virtual Fencing may be an effective way to create linear fuel breaks in rangelands

This trial suggests that VF is an effective solution to install linear fuel breaks in fire prone rangelands, and it will likely gain value as the technology continues to mature. This application seems most practical as an added benefit for ranchers and land managers who have invested in VF to improve other areas of their operation, such as monitoring and managing livestock across vast or difficult-to-fence areas, as is common in many summer ranges. In this scenario, VF collars could be deployed slightly earlier in the season to schedule fuel break installation in critical areas just before cattle are moved to summer pasture.

Want to learn more?
Reach out to Brian Allen at brallen@ucanr.edu

Or join our Livestock and Natural Resources Newsletter email list

Sign up to receive email news about future research

This work was funded by the USDA Natural Resources Conservation Service.

References
1) California Department of Forestry and Fire Protection (CAL FIRE). (2019–2023). Wildfire Activity Statistics Annual Reports. https://www.fire.ca.gov/our-impact/statistics

Virtual Fencing: Collaring and Training Livestock 

In an earlier article, we introduced virtual fencing (VF) as a system for containing livestock without traditional fences. Now, let’s explore how to collar and train livestock to recognize these virtual boundaries.

Setting Up Virtual Fencing Collars 

To set up collars, first, charge the collars and allow time for them to connect to your mobile app via cell signal. Once connected, you'll see their locations updating regularly on the app. The collars are now ready to be placed on the herd.

Fitting Virtual Fencing Collars on Livestock 

Properly fitted VF collars sit around the neck like a cowbell and are easy to attach using a traditional livestock squeeze. A properly fit collar should hang slightly so that the animal can breathe, eat, and drink freely. Being too tight may restrict this while being too loose may result in the collar falling off (Figure 1). Make sure the fit is checked regularly, especially if the animal’s weight changes.

Image

Figure 1. Example of a proper collar fit.

Training Livestock with Virtual Fencing Collars

Once collars are on, training begins! Place the herd in a pasture with a simple physical fence geometry, like a rectangle or square, and adequate cell reception (Figure 2). Provide easy access to water, feed, and shade that will be set at least 50 feet from the VF boundary. On the mobile app, draw the VF boundary so that one line cuts across the middle of the pasture. The pasture size will depend on the number of livestock but should allow them to roam freely while interacting with the VF boundary several times a day. Too little interaction won’t provide enough training, while too much may confuse them. For example, we trained 16 heifers in a 21 acre pasture, using VF to restrict it to 15 acres. Training usually takes 4 days to a week, depending on the livestock’s response. When you notice the herd staying within the VF area for at least 4 days in a row, they are considered trained and are ready to use VF in various ways!

Figure 2. Left: Example of a VF training pasture, with the barbed wire fence (white line) and only a single side of the VF boundary (orange line) crossing the middle of the pasture. Right: Location data (yellow points) of the herd after one week of training. The herd obeyed the VF boundary 99% of the time, however it is normal for some animals to cross the VF boundary during this period, as seen by the few location points in the southern part of the pasture.

Livestock response to VF cues

Livestock usually respond quickly and intuitively to the audio and electric pulse cues emitted from the collar when they cross a VF boundary. For example, in a trial with 37 cattle new to VF, they received the electric pulse about 26% of the time during the first week of training. As they started to understand the cues better, the need for electric pulses dropped to around 4% for the remainder of the trial, despite being rotated to different VF areas (Figure 3). This shows that the herd learned to respect the VF boundary primarily from the audio cues alone. Throughout the trial, the herd stayed within the boundary 99% of the time.

Figure 3. Daily percentage of audio (blue) and electric pulse (orange) cues from a herd of 37 cattle over a month. During the training period, the herd received electric pulses about 26% of the time. After the training period, the need for electric pulses fell to 4% on average for the rest of the trial.

Feel free to reach out to Brian Allen (brallen@ucanr.edu) if you’d like to discuss more about virtual fencing.

Learn more about Virtual Fencing

Virtual Fencing: A New Tool for Livestock and Land Management 

Virtual fencing (VF) is a precision agriculture tool gaining traction across California’s diverse rangelands. It uses GPS collars to contain livestock without physical fences. Producers set virtual boundaries across their pasture from a computer, which are transmitted to the collars. When an animal approaches the boundary, the collar emits audio cues and, if needed, mild electric pulses to keep the animal within the boundary.

Image

While not a replacement for a secure perimeter fence, VF can help position animals across a landscape with more intention and precision than traditional fencing. VF can be adjusted in minutes to meet the changing needs of livestock producers and range conservationists alike. 

Applications include:

• Targeted grazing to control invasive weeds
• Installing fuel breaks to reduce wildfire risk
• Automated rotational grazing
• Excluding livestock from sensitive areas
• Real-time livestock location tracking
 

With NRCS funding, the UC Cooperative Extension Central Sierra is researching VF applications on California rangelands. 

Contact Brian Allen (brallen@ucanr.edu) or visit our website to learn more!

Infestations of Medusahead can be suppressed by targeted grazing 

Medusahead (Elymus caput-medusae) is a common invasive grass on California’s annual rangelands. It accumulates high levels of silica as it matures, making it unpalatable to livestock and slow to decompose. As a result, livestock tend to avoid it for more desirable forage, potentially allowing it to form a dense thatch layer that persists on the landscape. This can suppress more desirable forage species and allows medusahead to spread over time.^1,2

Medusahead seeds remain viable for only a few years, so disrupting seed development can control infestations. While herbicides and mechanical removal are viable control methods, we explored whether high intensity, targeted grazing guided by virtual fencing (VF) could offer an effective and practical alternative.^1,2 Fortunately, there is a brief window when the developing plant has produced immature seeds and has not yet accumulated high silica levels.³ During this time cattle may find medusahead palatable enough to consume, potentially interrupting its reproductive cycle.⁴

UC Cooperative Extension Central Sierra Advisors used Virtual Fencing for targeted grazing 

In May of 2024, we used Virtual Fencing (VF) to concentrate 25 steers and heifers on a 3-acre Medusahead infestation along a fenceline within a larger pasture in Amador County. Water and shade were provided within the enclosure. The livestock stayed in the VF area 100% of the time for nine days. On the tenth day, however, the livestock left the area, presumably for better forage. Despite this escape at the end, we were satisfied with the result. 

Only 5.6% of the pasture in the target area still had ungrazed Medusahead with viable seed heads that could grow next season, whereas immediately adjacent to the grazing area, ungrazed Medusahead had turned into a thatch that covered 87.3% of the pasture (Figure 1). All animals gained weight during the trial.

Figure 1. Left: The viewer's left demonstrates the impact of high-intensity, short duration grazing on immature medusahead, contrasted with the ungrazed control area showing high levels of medusahead infestation, identifiable by its glossy yellow green color. The VF boundary is visualized by a line in the pasture that the animals are trained not to cross. Right: VF collar location data (blue dots) during the trial. The VF boundary is represented by a yellow line and the barbed wire fence is represented by a white line. 

Effects of high intensity targeted grazing using VF collar

The effects of this single high intensity targeted grazing treatment carried over into the following year. While the control plot remained nearly unchanged, the grazed plot displayed the following improvements one year later compared to the original condition:

• Medusahead cover remained far lower in the grazed plot, from 22% originally to 6%.
• Other grasses and broadleaf plants (forbs) doubled.
• Thatch cover was nearly eliminated.
• Bare ground remained low.

Image

Figure 2. Estimated percentage cover of vegetation types immediately before the grazing trial (blue) and one year after (gold). Left: Impacts of grazing on vegetation types. Right: Ungrazed control. 

Results of this Virtual Fencing grazing trial

These results suggest that virtual fencing has merit as a tool for targeted grazing of medusahead during its short vulnerable phase, offering ranchers and land managers a flexible and labor-saving option to improve rangeland health. Currently, we are conducting a similar trial on barbed goatgrass (Aegilops triuncialis), a similar problematic annual grass weed. We are excited to share those results with you in the future.

Want to learn more?
Reach out to Brian Allen at brallen@ucanr.edu
 

References
1.George, M. R. (1992). Ecology and management of medusahead. University of California Range Science Report, 23, 1-3.
2.Young, J. A. (1992). Ecology and management of medusahead (Taeniatherum caput-medusae ssp. asperum [Simk.] Melderis). The Great Basin Naturalist, 245-252.
3.Murphy, A. H., & Turner, D. (1959). A study of the germination of medusa-head seed.
4.DiTomaso, J. M., & Smith, B. S. (2012). Linking ecological principles to tools and strategies in an EBIPM program. Rangelands, 34(6), 30-34.

Tech can save ranchers time and benefit animals and land, becoming more viable

After the Caldor Fire destroyed seven miles of fencing on their cattle ranch in 2021, Leisel Finley and her family needed to replace the fence.

Finley, a sixth-generation rancher at Mount Echo Ranch in Amador County, said reconstruction costs were bid at $300,000 and would take at least a year to build, leaving the family without summer pasture and a herd of hungry cows to feed. Additionally, the U.S. Forest Service mandates that grazing be withheld for two years in postfire landscapes. This put the family in a difficult position.

While watching a recording of a California Cattlemen's Association meeting, Finley learned about a pilot program for virtual fencing. Desperate to find an alternative solution, she registered to try the livestock containment technology, which uses GPS enabled collars to monitor each animal's location in near real time.

Livestock producers can draw a perimeter on a map of their pasture using a laptop or smartphone application and send those instructions to the collar. The collar then uses audio and tactile cues to contain the animal in the area.

Eager to discover the short- and long-term benefits of virtual fencing, Finley turned to Scott Oneto, farm advisor, and Brian Allen, assistant specialist, from the University of California Cooperative Extension office in the Central Sierra. Since partnering with Oneto and Allen, Finley said she has come to understand and uncover more of the technology's potential.

The team has consistently observed the technology's value in integrating with and enhancing traditional livestock production systems across California. Though still in its early stages of development, the location tracking and containment system appears to provide time- and cost-savings that make it a game-changer for ranchers.

Ability to monitor location of animals in real time

Virtual fencing really stands out in its ability to monitor each animal's location in real time. During roundups, ranchers can use their smartphones to see their own location relative to their herd. The system can also send alerts if an animal crosses the virtual boundary or if a collar remains stationary for an extended period, potentially indicating that the animal is sick or that the collar has fallen off.

Rounding up cattle on large, forested grazing allotments can be challenging, as the process generally requires a group of people and many return trips to find every animal. Prior to virtual fencing, Finley and her father could gather about 85% to 90% of the herd in a week. Since using virtual fencing, Finley said one of their most recent roundups lasted three days, and they located every single cow.

Something that every livestock producer dreads is the notorious call from a neighbor or California Highway Patrol alerting them that one of their cows is out in the middle of the road. It always seems to happen at midnight or while they are out with friends or family. This scenario changes with virtual fencing.

Containment based on animal behavior

The containment system that virtual fencing is built on is based on animal behavior. When the animal crosses an invisible boundary, the collar emits an audio warning, prompting most animals to instinctively turn back into the desired area. If the animal doesn't respond, the collar delivers a mild electric pulse as a secondary deterrent.

Field trials by Oneto and Allen demonstrated the system's success. Recently, the team trained a herd of 37 cattle of mixed ages that had no previous exposure to virtual fencing. During the initial six-day training period, the cattle responded to the audio warning alone about 75% of the time when they approached a virtual fence boundary, with the remaining 25% of cases requiring an electric pulse.

After about three weeks, the herd was responding to audio cues alone about 95% of the time. The field trials also showed that the collars contain the livestock within the desired areas 90% to 99% of the time when the entire herd wears virtual fence collars and their basic needs for safety, connection to the rest of the herd, water, forage, shade, etc. are met.

Opportunities for improvement

While the technology is effective in its current capacity, there are notable areas where it can improve. One limitation to the system is the current reliance on cellular networks to operate. If an animal wanders into an area outside of coverage, the collar will continue to operate based on the last instructions but won't receive updates or report locations. This is especially a concern in many areas of California with poor cell reception, including the steep forested rangelands where many livestock producers have summer grazing allotments.

Another limitation is that some companies require a solar-powered base station with radio and cellular antennas to be placed on the pasture. These facilitate the transfer of animal locations and updates to the virtual fences. A base station going offline would create the same conditions as a drop in cell signal until the base station is repaired. Some companies are currently developing collars that bypass the need for these base stations.

The other major concern for ranchers is the cost for a virtual fencing system. The average rancher can expect to pay an estimated $20,000 to $30,000 in upfront costs. The cost to set up a base station alone is $5,000 to $10,000. However, this cost is highly dependent on several factors, including the manufacturer, the number of livestock to be collared, if the livestock are large or small ruminants, and the number of GPS base stations to cover the range.

According to Allen and Finley, the high cost of virtual fencing can be offset by the unique animal and land management benefits it can provide. “While physical perimeter fencing remains essential, VF is rapidly emerging as an innovative tool to control livestock with ease, precision, and flexibility in ways that were not previously feasible with traditional fencing,” Allen said.

Finley described the technology as a “game-changer” for her family.

Virtual fencing helps control invasive grasses, installing fuel breaks

While virtual fencing is designed to contain livestock without physical fencing, it is not intended to outright replace secure perimeter fencing. Instead, it operates best as a highly dynamic and adaptable cross-fence, allowing for more intentional grazing on the landscape to meet livestock production and natural resource conservation objectives within a secure physical perimeter.

With grant funding from the USDA Natural Resources Conservation Service, the UCCE team continues to work with Finley and other livestock producers to test these applications on California's diverse rangelands.

Within the Sierra Nevada and Coast Ranges foothills, these trials include using virtual fencing on cattle for targeted grazing of invasive grasses to support the recovery of native forage and installing fuel breaks within the wildland-urban interface to remove vegetation where the edge of a pasture meets urban housing.

Using virtual fencing, 25 cattle were successfully concentrated on a field of Medusahead (Elymus caput-medusae), an invasive annual grass. The herd respected the virtual fencing boundary 99% of the time despite nearby preferable forage. Grazing reduced medusahead seed heads from 2,072 per square meter in the ungrazed control area to just 68 per square meter in the grazed section.

In a different trial, 37 cattle with virtual fencing collars were contained within 120-feet-wide fuel breaks along the boundary of an annual rangeland and residential area. Cattle stayed within the boundaries 99% of the time, leading to an 81% reduction in fine fuel biomass and lowering wildfire risk in the wildland-urban interface.

Within rangelands on conifer forests, these UCCE trials concentrate cattle on brush to reduce the flammable plants and vegetation that competes with desirable timber species. It also can prevent livestock from entering sites that are sensitive to livestock presence.

Upcoming grazing trials will focus on how virtual fencing works with goats and sheep. In addition to Oneto and Allen, UCCE's contribution to virtual fencing research is in large part due to Leslie Roche, UCCE specialist and associate professor at UC Davis, Dan Macon and Jeff Stackhouse, UCCE livestock and natural resources advisors, Kristina Horback, associate professor at UC Davis and Lone Star Ranch in Humboldt County.

To learn more about the trials led by the UCCE team,visit https://cecentralsierra.ucanr.edu/Virtual_Fencing/