Testing virtual fence systems for fire fuel management

Results

Effective Containment

The results of the trial demonstrated that the collars did not contain all of the animals precisely in the boundaries for the duration required to achieve appropriate impact levels, as some animals repeatedly moved out of the boundaries. However, we were still able to achieve effective impact as most of the animals did remain contained and created a desire for those who broke the boundaries to return to the larger group.

Our conclusion is that the collars we tested could only be used in a narrow set of circumstances to contain animals for fire fuels management. Containment was not effective on a portion of the animals consistently enough and the majority of fire fuels work would require stricter adherence to the system.

We conclude that containment was affected by several factors that could improve results using the Nofence system. These factors are:

• Connectivity issues
• Wool breeds
• Density of grazing and area of the paddocks
• Poor behavior enforcement

Connectivity: There were many instances when the collars did not connect to the system, recognize freshly charged batteries, or had issues recognizing precisely where the collar was. We believe a portion of the collars could have not been properly recognizing where the animals were, or were not moved to the new paddocks and therefore there were no consequences to those animals who left. This was a minor condition—usually only one or two collars in use would not connect properly—however these animals would not respect the boundaries and encourage other animals to break containment.

Wool Breeds: A factor we cannot measure but might have affected the effectiveness of the shock on the sheep was the growth of wool. The animals were trained to the system when their wool was recently shorn, but as it grew in, it could be that they were not receiving as much of a shock as they were initially. This is a speculation given that we only tested wool breeds, but we have heard from people using goats and hair sheep stories of more effective restraint. That being said, the majority of the wool sheep were effectively contained consistently, so behavior of individuals may have had more of an impact on containment than breed. Wool sheep should not be disregarded as a potential for application of this technology without further trials.

Density: In the density mimicking trial the area was so small that the sheep were constantly bumping against the boundaries. This resulted in a lot of warning sounds and could have desensitized the animals to the warnings. Certain animals became unconcerned with the boundaries and continually broke out. The number of animals who would break containment was small at first, but in later parts of the trial there were over 10 animals a day breaking containment regularly. You can see from the screen shots how breaking containment affected the other animals who would gather on the boundary and try to follow the other sheep, encouraging others to take the risk and cross the boundary line. As the trial went on in that small paddock, more and more animals proceeded to escape more regularly. It also meant animals were shocked more often and the collar batteries were being drained much more quickly. We had to replace 30% of the batteries after that short trial, and could have benefited from replacing more. Larger paddock areas were much more effective at keeping animals contained and the lower the feed in a given area the more pressure there was for animals to escape. In a project context that did not require 70% reduction of fuels, the collars also could have been much more effective.

Behavior: All of these factors contributed and increased the levels of poor behavior in the sheep and reinforced risk-taking and disregard for the system. As animals began to repeatedly break containment we saw more animals take the chance to join them. Even in electric fence systems if one or two animals begin to escape, the pressure of them being outside of the boundaries can lead animals to push through fence and leave containment. We also saw as the group that was escaping became larger they were more comfortable wandering further from the main group. It is unknown if more animals would continue to escape or if there were a certain number of animals who had the tenacity to disregard the system and the number of escapees would cap somewhere based on the character of the animals. It is also unknown if a retraining period, shearing, or replacing collars for the animals who were consistently breaking containment would reinforce the system and bring animals back to regular good behavior.

Because the trials took place in areas where brief escapes weren’t a concern, the level of containment we achieved was sufficient and the animals still made a meaningful impact on the vegetation. However, the amount of time they spent outside the paddocks significantly reduced the effective grazing rate. In a contract grazing scenario, this would lower both the quality of work completed and the pay received when compensation is based on dollars per acre grazed. After initially responding when animals broke out we soon stopped going to return the animals to the area and allowed them to go back on their own, which meant, despite breaks of containment the system was very low management during that time. Furthermore, the GPS system allowed me to see on my phone where escaped animals were and know I would be able to easily recover animals that wandered too far.

In conclusion, despite a lack of reliable containment of all of the animals, the collar system could be used in very specific fire fuel contexts where strict containment is not a large priority, like in remote areas and areas where electric fencing is not possible.

Labor Comparison

The labor required to run each system was broken down into two components: labor input for building and dismantling paddocks and labor for maintaining the system, which included replacing batteries in the Nofence system and checking the electric fence for any issues.

We experienced robust savings in labor for building and dismantling paddocks using the Nofence system (Table 1a). During our trial period, electric fence Paddock 1 took a total of 18 labor hours to build and dismantle while Paddock 2 took 15.5 hours to build and dismantle. Labor to build the paddocks included: dropping net, putting net up, adding posts to the section to reinforce the electric netting, and setting up the energizer. Dismantling the paddock included: removing posts, rolling up nets, and compiling nets.

The Nofence paddocks took only minutes to set up, and even less time to “take down”. Building paddocks involved: looking at the site where the desired paddock would be and tracing an area out with your finger on the Nofence app, following any helpful natural contours or boundaries in the area to shape a paddock that makes sense to the livestock, and ensuring that there was ample space around the water trough. Dismantling the paddock involved nothing more than moving the sheep out of the previous paddock.

In total, labor for building and dismantling paddocks during the trial took 33.5 hours for the electric fence system and .22 hours for the Nofence system.

Table 1a. Labor Comparison - Building and Dismantling Paddocks (Trial Period)

We extrapolated the labor hours required during the trial period for a hypothetical year (Table 1b). These numbers are highly speculative as the amount of time required to build paddocks can vary widely. There are instances where weed whipping is required and occasionally even chainsaw work, though we did no weed whipping for the construction of the trial paddocks. At other times we can follow preexisting paths in a park, thus lessening the labor required. For a group of 400 sheep, we extrapolated that they may require 80 paddocks per targeted grazing season. Given the labor hours we clocked during the trial period, we estimated very roughly that an average paddock would take 10 hours to build and 10 hours to dismantle, and require an additional 10% of construction time to troubleshoot. The total hour estimate for building and dismantling paddocks for a full season of targeted grazing totaled 1,680 hours. In contrast, the Nofence system, at .1 hours per paddock, would take approximately 8 hours of labor to build 80 paddocks, and under an hour to "dismantle" those paddocks. We added an additional ~7 hours for troubleshooting paddocks on the app (adjusting boundaries according to real life conditions, for example, or fixing the proximity to sensitive areas or the water trough). The total for the Nofence virtual fence system totaled 16 hours, constituting an annual reduction of 1,664 labor hours compared with the electric fence system. This constitutes by far the largest savings in labor hours, and thus cost, in the Nofence system.

Table 1b. Labor Comparison - Building and Dismantling Paddocks (Seasonal Estimate)

Labor for maintaining each system included the time it took to charge batteries, build and dismantle a corral to work the Nofence flock in, and the time it took to swap out batteries from the collars and troubleshoot any resulting issues. Initial setup and installation of the collars took 12.5 hours for a total of 61 sheep. Extrapolating to a flock of 400 sheep we considered the total time could take approximately 49 hours (certain aspects we did not think would increase linearly with the increase in number of sheep). 

Subsequent labor for maintenance of the virtual fence system included charging batteries, corral set-up and take down, and swapping out batteries. We did this twice during the remainder of the 10 week trial, however the short lifespan of our batteries really suggested that doing more battery swap outs would have benefited the maintenance of the system. With improved battery life and/or more batteries on hand we could have replaced all batteries under 70% charge (for example) rather than under 30% charge (which is what occurred in practice). This was one factor that contributed to behavioral issues with sheep not respecting boundary lines—a collar running out of battery and a sheep leaving the boundaried area, emboldening other sheep to do the same. We made an informed estimate of what the labor needs would look like for a flock of 400 sheep (a typical targeted grazing group).

Table 1c. Labor Comparison - Maintenance of System (During Trial)

Table 1d. Labor Comparison - Maintenance of System (Year Estimate)

For a flock of 400 sheep equipped with Nofence collars we estimated that maintenance of the system would require approximately 72 hours for the trial period (10 weeks) and thus 374.4 hours for the year (Table 1d).

Maintenance of the electric fence system was estimated at a modest .25 hours per day. This number is highly variable, with many days when no particular maintenance is required, and other days when a couple of hours of maintenance (walking the line, fixing breaks, finding ground-out spots) is required. At .25 hours per day, 91 hours per year would be required to maintain electric fence systems, a difference of 283.4 hours, or 35.4 8 hour work days.

Projected Costs

Initial Cost of Investment

We determined that the initial cost (based on current pricing as of this report) would result in a nearly $100,000 difference in start up investment between the Nofence virtual fencing system vs. a typical electric fence system for a group of 400 sheep in a targeted grazing system. Certain add-ons, such as the leather collars with the resulting need for extra chain lengths versus the silicone collars that come with each unit, are optional depending on conditions, but were added due to issues with grass seed we experienced on our wool sheep. The number of batteries and chargers listed could be modified, particularly with newer battery life improvements that have been made on models newer than the 2.2 model we utilized. However the bulk of the cost comes directly from the collars themselves.

Table 2. Cost of Initial Investment

Annual Cost of Investment

The major difference in annual costs is in the annual subscription required to run the Nofence system. Though there will be significant variability in the annual costs for the infrastructure of each system, the annual subscription is a reliable due that needs to be paid for the virtual fencing system and outstrips any probable infrastructure cost of the electric fence system.

Table 3a. Comparison of Annual Infrastructure Costs: Electric Fence System

Table 3b. Comparison of Annual Infrastructure Costs: Nofence Virtual Fence System

Comparison in Cost of Labor

The savings in labor time and cost is definitely where the Nofence system excels. With electric fence, as the size of paddocks increases and the terrain becomes more challenging, the paddocks take longer and longer to build. Sometimes lines need to be cut with chainsaws through brush. Often weed-whipping prior to building lines is essential. Paddocks need to be built and then reinforced with posts, increasing the labor requirements in building and dismantling paddocks. With the Nofence virtual fencing system, none of these factors increase the time it takes to build or dismantle a paddock.

In this trial the electric fence paddocks we built were relatively straightforward compared with many areas we work, though they were on the large side. Table 4a compares the cost of labor for building and dismantling paddocks during our trial period for each of the fencing systems. Table 4b extrapolates those numbers for a full targeted grazing season.

Table 4a. Cost Comparison of Labor Needs - Building Paddocks (Trial Period)

Table 4b. Cost Comparison of Labor Needs - Building Paddocks (Year Estimate)

To estimate the labor required for the maintenance of each system, we extrapolated the time of maintenance during our trial to deduce the amount of time maintenance might take throughout the year (Table 5). Many factors could affect this number: the weather most of all can lead to battery drainage much more in the winter than the summer, as well as factors contributing to sheep running into the edges of the paddocks and receiving alerts, thus contributing to the draining of batteries. This is more likely to occur in small and dense paddocks. The biggest factor of all is the length of battery life, which is improved in more recent iterations of the Nofence system (2.5 and later). Our trial ran in the summer months, and even so we saw a frequent draining of batteries.

By our estimates the Nofence system requires about $7,000 more in maintenance annually than the electric fence system. Again, these numbers are difficult to estimate as there are many variables that can affect them.

Table 5. Cost Comparison of Labor Needs - Maintenance of System (Annual Estimate) 

Total Cost Comparison

Table 6 summarizes all the measured costs of the system and projects them out 1, 3 and 5 years. This projection is based on a group of 400 sheep, the number of sheep required for a typical targeted grazing project, but does not necessarily reflect the actual size of a typical operator. After the first year of initial investment and one year of labor costs the Nofence system costs the operator $80,680 more for their group of 400 sheep. After 3 years, due to the lower annual labor costs of the Nofence system, the Nofence system has cost the operator a total of $27,530 more than the electric fence system. After 5 years the operator experiences a net savings of $21,760 utilizing the Nofence system. Each year the Nofence system saves over $40,000 in labor costs in our projection, which means the system for 400 sheep equalizes in cost after approximately 4 years of use, after which it provides a net savings to the operator.

Table 6. Total Cost Comparison

These numbers are informed and yet highly theoretical. Some of the factors that would dramatically change these calculations could be: having an increased battery life for the Nofence collars (reducing cost of the virtual fence system); having many paddocks that require on average less labor than our estimate (reducing the cost of the electric fence system); having more than 400 head (increasing the initial investment of the Nofence system); or receiving a grant to support the initial investment in the collars (reducing cost of the virtual fence system). 

Additional Discussions

LGDs

Using a virtual fencing system successfully requires the operator to have a guardian animal that stays with the flock, either by instinct and training or utilizing a virtual fencing system suited for that species of animal. We use Livestock Guardian Dogs (LGDs) in our operation who are accustomed to electric fence and who do not stay with the flock without that fence. We ended up borrowing dogs from another operator who are used to working with their flock without the presence of fence. Another option could be to use donkeys, who could be trained to the Nofence collars in the same way as the livestock.

Public Areas

Targeted grazing often occurs in areas known as WUI zones (Wildland Urban Interface). It is not uncommon for these areas to abut neighborhoods, or be open to the public with public access trails running through. In these instances the electric fence serves not only to keep the livestock in but also to keep the public (and particularly their dogs) out. In addition the electric fence provides a precise barrier to sensitive areas, such as landscaped yards, whereas the Nofence virtual fencing provides more of a boundary zone of 5-10 yards, which may not protect sensitive areas as precisely, or would sacrifice the effectiveness of grazing impact near the boundaries.

In addition, in many public areas it would not be safe to have an LGD present with the livestock, even one who stayed with their flock. If they perceive a danger they may well go after that person/dog/bicycle without being enclosed in fence.

Grazing areas that are more remote or on private land are ideal for employing the virtual fencing system.

Predation

We experienced one attack during the span of our trial. It appeared that the attack was from a visiting dog, not a coyote. This attack would not have occurred if the animals were secured inside an electric fence system.

Conclusions

Across the three areas evaluated—effective containment, labor inputs, and costs—the Nofence virtual fencing system showed a mix of advantages and significant limitations.

Containment:
The collars did not reliably contain all animals within the virtual boundaries, particularly under conditions of high grazing pressure, high density, and with connectivity issues and battery failures, all leading to behavioral contagion among sheep. These factors contributed to repeated boundary breaches by a subset of animals, reducing grazing efficiency and reinforcing poor behavior in the flock. Nonetheless, most animals remained contained most of the time, and in low-risk, remote settings the system still enabled meaningful vegetation impact with very low active management. For fire-fuels projects that do not require strict containment or precise boundary protection, the system could be suitable.

Labor:
Labor demands between the two systems differed dramatically. Virtual fencing required almost no labor to build or dismantle paddocks and relatively modest ongoing maintenance—primarily battery charging, corral setup for battery swaps, battery swapping, and troubleshooting. In contrast, physical fencing required extensive labor to construct, reinforce, and dismantle paddocks, especially in varied terrain or brush-heavy areas. Over a typical grazing season, the virtual fence system reduced labor by an estimated 1,664 hours, representing the largest operational advantage of the technology.

Costs:
The cost comparison showed substantial tradeoffs. Virtual fencing required a far higher initial investment, primarily due to the collar purchase and subscription fees. Annual operating costs were also higher because of the cost of subscription. However, these costs were partially offset by significant labor savings. Based on projected annual labor and equipment costs, the system is expected to reach cost parity with electric fencing after roughly four years, after which it may offer net savings—assuming battery performance improves and labor estimates hold.

Overall:
The Nofence system is not yet a universal replacement for electric fencing in targeted grazing, especially where strict containment or public-interface safety are required. However, in the right conditions—remote sites, low public risk, minimal need for exact fencing lines—it offers considerable labor efficiencies and long-term cost potential. As the technology continues to develop, especially regarding battery life, connectivity, and training efficacy, its applicability is likely to expand.