Evaluating virtual fences for cattle in regards to water resources, forage management, invasive weed control, and wildlife systems

Progress report for LNC22-476

Project Type: Research and Education
Funds awarded in 2022: $249,999.00
Projected End Date: 10/31/2025
Grant Recipient: Kansas State University
Region: North Central
State: Kansas
Project Coordinator:
Jaymelynn Farney
Kansas State University
Expand All

Project Information

Summary:

Determine feasibility of virtual fences on cattle operations to reduce labor costs and improve/sustain water and pasture resources with flexible grazing management. A new technology that is becoming available for cattle producers that can help with a wide range of environmental, ecosystem, and economic decisions on their operation is virtual fences.  Many production practices that improve ecosystem and natural resource management are not implemented in production because of lack of capital, labor, topography concerns, and an ever-changing environment.  We propose to test virtual fences on cattle operations for a variety of established natural resource management systems.  Four research projects will be conducted while also determining the viability of the technology in a hilly, hard-wood terrain.  Study 1: Evaluating virtual fences to develop point-specific watering sources within riparian areas or around ponds.  We propose to test the effectiveness of virtual fences to improve/maintain high quality water and maintain bank integrity by implementing limited access watering sources. Study 2: Utilizing virtual fences to implement ultra-high stocking on areas of pastures infested with the noxious plant Serecia lespedeza.  This novel management practice will be compared to traditional management practice of herbicide treatment, a negative control with no mitigation, and a new practice of fall-burning pasture.  Study 3: Evaluate the use of virtual fences to manage stocking rate and grazing pressure in songbird nesting habitats develop cohesive grazing and wildlife management operations.  Study 4: Evaluate the use of virtual fences to aid in removal of livestock from croplands during periods of high moisture.  This study will also evaluate the effects of grazing through the high-moisture events and no grazing of the cropland on subsequent crop production.  The information attained through these projects will be made available to three main audiences – scientific community, livestock and land-manager audiences, and government agencies.  Outreach efforts for all three groups will include peer-review journal articles, presentations at professional meetings, field days, portable virtual fence demonstration, popular press articles, media interviews, podcasts, and highlights through social media platforms.  A working group with government agencies to discuss feasibility of virtual fences as a cost-share program will be conducted along with special tours for these groups.  Assuming the virtual fence technology works, there are potentially limitless opportunities for producers to be able to manage their resources to the micro-degree while enhancing quality of life for the producer, area, and will provide much needed quick adaptability to a rapidly changing environment.

Project Objectives:
  • Overall objective: utilize technology to incorporate management practices that improve natural resource in constantly changing environments or hard to manage areas.
  • Sub-objectives: 1) manage for water resources/ponds; 2) improve habitat for wildlife species of concern; 3) improve pasture management via production and/or invasive weed control; and 4) fencing assistance in integrated cattle-cropping systems.
  • Action outcomes: technology will lead to use of strategic grazing systems that have been proven to enhance natural resources, but have not been incorporated due to lack of labor, topographical feasibility, and constantly changing environmental conditions.
  • Audience: scientific community, cattle producers and stewards/managers of land, government agencies
Introduction:

Best management practices (BMPs) have been identified for many different factors to manage ranch water quality and environmental performance, especially as related to cattle production.  All these management strategies take large amounts of labor, capital investment, and may not be feasible based on topographic pasture profiles.  Unfortunately, labor is one of the major issues with cattle production. Many of these BMPs utilize fencing on the property.  Utilizing permanent cross-fencing on owned property can result in an overall improvement in land capital, however, many cattle operations utilize rented property, in which cattle producers may be reluctant to invest capital into that property. Fencing riparian areas, especially, poses a financial and logistical challenge to farmers and ranchers, as fences are very expensive ($5000/mile or more) to install, can impede cattle and wildlife movement within and between pastures, and are difficult to install over drainages and uneven topography, and require significant labor to maintain.   

 

Which BMPs rely heavily on fencing? 

Riparian area (area around streams) and pasture water source management. Some commonly accepted BMPs for water management include stream crossings, rotational grazing, pasture improvement, and cattle water tanks.  These are areas that have been promoted as functional to reduce sediment, nutrient, and fecal coliform contamination from cattle.  However, there are varying levels of acceptability by producers to these management practices.  A survey from Tennessee found that physical and economic constraints and financial incentives from state and federal programs are primary drivers of producer’s acceptability (Lambert et al., 2014).   

Riparian areas are attractive to cattle because of shade, predator protection, and water (Bryant, 1982).  Due to these factors, riparian areas are often overgrazed and cattle spend significant time around the water source, potentially leading to water contamination.  Completely removing cattle from riparian areas is not the solution, as there are some benefits to grazing riparian areas, such as protection from woody overgrowth and creation of new wildlife habitats (Larsen et al., 1998).  Developing a system for strategic grazing/non-grazing, riparian fence widths, and water development systems will become important for managing this resource from a grazing perspective.  Even though off-stream watering points have been effective at managing water quality, a meta-analysis illustrated that there are many factors that impact the effectiveness of this practice such as slope, pasture/paddock size, shade, temperature-humidity index, and distance to stream (Malan et al., 2018).  However, these authors still felt that there was not sufficient data to develop guidelines for location of off-site watering points that would be specific for producers and locations.   

Strategic grazing for weed control. Serecia lespedeza is a poor forage crop has it is high in concentrated tannins; however, calves will consume this plant when it is immature and tannin concentrations are relatively small (Wang et al., 2008).  Providing grazing pressure, as a biological control to Serecia lespedeza may offer a chance to manage for this invasive weed.  Goats have a higher tolerance to tannins than cattle (Hart, 2001) and at high stocking rates were able to control Sereciea lespedeza.  Utilizing high intensity-low frequency rotational grazing was more effective at reducing stands of Canada thistle than a short duration, low-frequency rotational grazing system and much more than a continuous/season-long grazing (De Bruijn and Bork, 2006).  Being able to utilize a biologic control for weed species provide an economic and ecologic benefit, especially as being able to reduce herbicide application. 

Grazing management for wildlife habitat maintenance or development. Grazing and pasture management methods can impact multiple wildlife species.  In a study using Northern mixed-grass prairies that are nesting locations for songbirds, researchers found that twice a year rotational grazing events resulted in 2.4 to 4 times lower nesting success for obligate grassland bird species, but no effects detected for shrub-nesting species (Carnochan et al., 2018).  Rest-rotation systems is a grazing management plan where individual range units are not grazed at the same time each year and are rested the remainder of the time.  Rest-rotation has been proposed to provide better grouse habitats than continuous grazing, even though with moderate stocking rates there was minimal effects on grouse nesting and survivability (Milligan et al., 2020).  Potentially, the selective foraging of cattle can provide the differing vegetative height requirements for grouse population coupled with topographical and edaphic variation naturally creates the necessary heterogeneity within pasture for grouse habitat, regardless of grazing system. There can be a beneficial relationship with cattle grazing pastures in regards to nest survivability, potentially through predator behavior which might have been reduced as cow-calf pairs were used in the pastures (Kluever et al., 2008). 

Grazing cropland by cattle. Cattle grazing on wet soils can lead to pugging, minor areas of compaction, and reduction in soil structure and thus producers grazing crop-ground would like to remove cattle from fields during periods of high moisture.  One of the historical studies that evaluated the effect of cattle on wet ground was conducted in Texas were bare ground was “wet” and varying stocking rates of cattle were placed on the pastures then soil hydrologic characteristics were evaluated (Warren et al., 1986).  In the study they found that the heavier the stocking rate on the wet soils the more deleterious effects to the soil were observed, and this hydrologic negative effect were not recovered even after 30 days of rest.  During wet events while grazing cover crops in Nebraska, soil compaction did occur in tracked areas which reduced water infiltration and potentially more runoff and erosion in those areas (Lesoing et al., 1997).  There are some management practices that can help manage for potential compaction effects, especially by not grazing in wet conditions, grazing when soils are dry and frozen (Clark et al., 2004).   

  

What is virtual fencing? 

Livestock virtual fencing companies have only been in business since 2005 and most of the companies are still in beta testing of the product.  The concept behind virtual fencing is that a cow wears a collar that connects with a transmitter/tower and GPS location to set specific boundaries for livestock (Umstatter et al., 2015).  The collar initially sends a warning sound to the animal as they approach the boundaries of the grazing area, then if the animal continues a slight electric shock will be applied.  The first “invisible fence” was patented in 1971 and primarily developed for dogs and cats.  Then goats were used with this technology and finally cattle were tested (Monod et al., 2009).  The first virtual fences contained a collar, battery-based transformer, and induction cable laid on or buried in the ground that marked the allowable location for the animal.  With improvements in technology, GPS has been incorporated to allow for larger areas of grazing and easier ability to modify grazing pastures and to record animal behavior.  In initial beta tests, cattle have adapted and learned this technology, in small research-controlled experiments.  A study in Australia found that utilizing virtual fence worked to keep cattle from riparian areas in a short-duration study (kept cattle off riparian area for 10 days, allowed access for 6 days) and the cattle were well contained with this technology (Campbell et al., 2018).  Heifers were able to learn to rotate to different areas of pasture with virtual fence (Campbell et al., 2017) and virtual fences have been identified to work as well as electric tape fencing for a period up to 4 weeks without substantial behavioral and welfare impacts on cattle (Campbell et al., 2019).  Published studies evaluating this product have not been conducted in the United States. Now a company within the United States is developing and beta testing with large herds within the United States (Vence Corp, San Diego, CA). 

 

How might virtual fencing be used to help with ranch and farm management? 

Flexibility of grazing is the proposed method to improve ranch and farm management of natural resources.  Virtual fences have the potential to allow cattle producers to be able to be more concise and deliberate with their management strategies.  As the producers that are involved in this project have mentioned, this technology can allow them to free up an employee to do other work, instead of just being hired to manage fences and watering sources. 

Research

Hypothesis:

Hypothesis 1: Virtual fences can be used to develop limited access water sources without having to build permanent fences.

Hypothesis 2: Virtual fences will make it easier to rotationally graze which can improve bird populations on ranches.

Hypothesis 3: Virtual fences can be used to quickly concentrate cattle to areas were undesirable weed populations are and increase intake of these weeds to potentially decrease their persistence in pastures.

Hypothesis 4: Virtual fences can be used in managing cattle on cropland.

Hypothesis 5: Virtual fences can reduce labor needs, improve operational management, and improve ecosystem management systems.

Materials and methods:

All the projects listed will be completed on the G3 Cattle operation in Uniontown, KS.  The virtual collars will be placed in a working cow-calf and stocker enterprise.  The owners, manager, employees, and former WaterShed Specialist have identified management issues they would like to address on property and how virtual fences may help them with management practices.  General care of the animals and using the GPS location data will be used by the researchers, extension specialists, and members of the G3 Cattle operation to help with future management decisions and areas of research.  The virtual fencing system is cloud-based so multiple members can use the data to track location of cattle and be able to evaluate the effectiveness of the grazing parameters being built for the cattle operation. 

 

Cows (n = 500) will wear collars throughout the entire 3-year research and observation period and as such will be easy to assign to different research projects.  When not used in designated pastures or fields for the research projects, the collars will be tested in a variety of environments to determine the number of animals that “strayed” from the designated fenced area.  This will help with determining the success of the equipment.  The producers will also look at animal location and overlay GIS maps to be able to use the information to help evaluate grazing distributions within pastures so that within a few years, a realistic rotational grazing paddock system can be developed using this technology.

 

Based on discussions with the operators at G3 Cattle Company, there are 4 areas of interest to evaluate whether the virtual fences can help with certain management practices and have been assigned to 4 research projects.

 

Project 1: Using virtual fence to limit access to pond and/or streambanks.  Virtual fences will be used to develop water access points for cattle with the objective of reducing streambank/pond side erosion and limiting access to the entire water source to reduce contamination of watering sites.  A minimum of 3 pastures will be assigned to the treatment of “free-access to water sources” and another 3 pastures will have designated drinking locations within the pond/stream.  The designated drinking locations are using the model of limited-access watering locations where traditionally a permanent fence is built around the watering source and with one designated area to drink water.  Ariel and ground photographs, GPS, and permanent markers will be used to identify initial streambank or pond edge locations.  Every year photographs, GPS locations and measurements from basal permanent marker will be used to measure erosion and plant species changes in the immediate riparian area and/or pond.  Ponds will be sampled yearly to determine water quality based on treatment imposed.  Water quality measurements will include nutrients and contaminants of concern for livestock species and total coliform and E. coli counts.

 

Project 2: Using intensive grazing via virtual fence during important plant maturity of invasive species, Serecia lespedeza, to reduce plant populations in pasture.  Three locations within ranch that are heavily infested with S. lespedeza will be measured for plant populations utilizing permanently identified transects and assigned to 1 of 4 treatments.  Treatment 1. Negative control – no management implemented; 2. Positive control - spraying with herbicide for control in spring; 3. Intensive grazing using virtual fence with cattle during spring and fall to minimize seed production.; and 4. Fall pasture burning (burn in late August or early September).  Treatments will be applied for 3 years.  Areas that treatment will be applied will be variable in size as the negative control, positive control, and burning areas will only need to be 3-meter x 110-meter areas.  The virtual fencing treatments will be a minimum of 10 acres (4.0 hectare).  Cattle will be excluded from the areas that are enrolled in treatments 1, 2, and 4 for 45 days after herbicide treatment has been applied to remove any potential effect of grazing during the spring grazing period.  Exclusion will be done by utilizing both virtual fences and solar powered electric fences.  Yearly plant populations, composition, and forage quality within treatments will measured in the fall.  Plant population and composition will be measured using a modified step-point method in the fall of each year.  Briefly, for the modified step-point method a 100-meter permanent transect will be placed within each treatment and every meter a pointed instrument will be dropped on the “line” and the plant that the point is touching will be identified.  Additionally, the nearest legume or broadleaf will be recorded.  In the fall of each year, 3 locations within each treatment will be harvested (1 ft x 1 ft) and seed will be collected from several locations in the treatment groups to count for total seed production based on management practice.   

 

Project 3: Strategic grazing to improve habitat for grassland songbirds.  Virtual fences will be used to promote a shifting mosaic of vegetation structure and composition by applying intense grazing to some areas and light to no grazing in other areas throughout the ranch.  The following year, treatments will “rotate” so that intensively grazed areas from the previous growing season will rest.  This system should provide heterogeneity that is currently lacking from many ranches in the Great Plains, and is crucial for imperiled songbirds, such as the Henslow’s sparrow (Centronyx henslowii).  We will measure vegetative response (habitat) to grazing treatments throughout the growing season, estimate grassland songbird densities using point counts, and determine nest survival by locating and monitoring songbird nests. 

 

Habitat measurements will include percent cover of plant functional groups, litter, bare ground, litter depth measured with a Daubenmire frame, and visual obstruction measured with a Robel pole.  Permanent habitat transects will be established in each treatment area, and will be sampled during the early, mid and late growing season for three samples each year.  Transects will be established in a stratified random design within treatment areas and will be located at least 100 meters from treatment edges to avoid boundary effects.  Moreover, habitat samples will be collected at each songbird nest along with a paired random point located within 100 meters of the nest.

 

Point count surveys will be conducted at equidistant points along transects established for quantifying habitat responses of treatments.  Surveys will start at sunrise and be completed before 11:00 with winds less than 35 km/hr and when it is not raining or foggy.  All birds seen or heard at each point will be identified to species, and we will record the perpendicular distance to each individual for density estimation via distance sampling methodology. 

 

We will systematically search treatment areas for grassland songbird nests via rope dragging and will opportunistically locate additional Dickcissel nests by observing behavior of mated pairs.  Nests will be marked with a painted rock positioned 5 m directly east of each nest and will be observed every 2-3 days until fledged or failed.  At each visit we will count all eggs and chicks to determine hatching and fledging success, as well as potential brood parasitism by brown-headed cowbirds. We will determine the cause of failure by examining eggshells and other signs at the nest site.

 

Project 4: Using virtual fences to remove cattle from fields in integrated crop-livestock systems during moisture events.  Research has found that excessive periods of moisture can lead to issues with compaction in crop fields, especially in higher clay soils.  We propose to evaluate the effectiveness of using virtual fences as a temporary method to remove cattle during high moisture weather events to maintain soil/field integrity.  Treatments are control (no removal of cattle during grazing period); removal treatment (remove cattle when moisture event (rain or snow) results in more than 0.5” occurs); and no grazing.  Replications of each treatment will occur.  Cattle in the removal group will be “penned” virtually with hay at the edge of crop field or into a perennial pasture until field is considered dry enough to continue grazing.  The “penned area will rotate to different edges of the field to avoid soil erosion.  Cattle will be grazing a winter annual forage according to ranches operational system.  Cattle will be turned out when the forage is appropriate to support cattle intake and plant is well established.  Grazing will end either a month prior to establishment of next crop, spring moisture is leading to severe issues in the grazing treatment, or the forage has been depleted and will not support livestock or provide residual to maintain soil health attributes.  Soil compaction and soil nutrients will be measured prior to grazing and after grazing for two years.  Forage biomass before grazing and after grazing will be measured and cow and/or calf weights will be measured. Subsequent crop yields will be recorded for each treatment replication.

 

Cost of production and economic feasibility of the technology in the systems will be accessed and incorporated into all the publications related to the research projects.

 

The owners, manager, and employees of G3 cattle company will be involved in all the decisions about treatment layout and applicability within the working ranch.  One graduate student will work in all the research projects and thus will be a multi-disciplinary trained student as they will be become well versed in understanding technology, reading GPS maps, forage management, cattle care, wildlife management, and understanding of all the aspects that are involved in maintaining a profitable, environmentally stable cattle operation.  Additionally, several undergraduate students and visiting students will be working on portions of the sample collections.  Working groups and tours with NRCS, watershed specialists, and industry groups are planned to help demonstrate the applicability of this technology for cattle operations. 

 

Additional dissemination methods will include peer-review journal articles for cattle, forage, wildlife, and cropping systems researchers and professionals within those respective industries.  Presentations about the projects will be given at national and regional agronomic professional meetings and to extension agents.  Information about the research will be presented at several field days that are primarily attended by cattle and forage producers, but also have industry professionals in attendance.  Periodic updates about the technology and management practices under investigation will be posted on social media platforms and documented and shared via videos, podcasts, radio, and television outlets.

 

In collaboration with other industry and government agencies, a survey about Kansas producer usage of best management practices (BMPs) and comfort with technology tools will be offered in a virtual and hard-copy form.  The survey will be used to determine future areas of investigation and the ability of virtual fences to be incorporated in increasing BMP usage.  The information collected will also be useful in discussions with government agencies about the feasibility of using virtual fences to implement known BMPs and utilizing cost-share capabilities.  Additionally, determining the comfort of producers with technology will be useful to provide a baseline to producer acceptance of the virtual fence technology.

Research results and discussion:

Project 1: The hypothesis and test for the objective (#1) where we wanted to determine if the virtual fences could work as a method to control water quality by restricting animals to a specific area of the pond, stream, or creek, DID NOT work.  We feel confident in telling producers with the current technology tested, that this is not a tool that can be useful in their ecosystem management practices.  The issue with the current technology in the extremely hilly and hard wooded area where we were located is that there is was a greater than 15 meter (~45 feet) margin of error with location of the animal and the "fence".  For ponds, when including that margin of error we had to have an opening for drinking that was over 60 meters (~180 feet) to guarantee only 15 meters (~45 feet) of water access, consistently.  Unfortunately, the cows we used were so well trained that when they heard the audio sound they immediately backed up and we had some cattle that did not want to try again.  Since water in such a vital nutrient, when we found those cows that were extremely timid and not making it to the water source, we shut the collars off for those animals and they were able to drink anywhere, which defeated the purpose.  Because we deemed a 180 foot (60 meter) area that may or may-not allow cattle access to drink as too large and inconsistent in allowing animals to drink, that this currently is not a viable option.  With improvements in technology and more accurate location of the animal and virtual fences this may become an option, but to-date it is not.

Project 2: Due to the extreme drought experienced through the 2022 and 2023, the pastures on this operation where the virtual fences were located were grazed so hard that the actual plant, S. lespedeza was not able to be found in sufficient quantities to test this practice.  The research team is currently working on developing a portable tower system to be able to test this hypothesis on other operations.

Interestingly, an informal survey of producers in the extreme drought area, found that there was a significant reduction in the population of the noxious weed, Serecia lespedeza.  This plant was the only thing early in the summer season that was green and as such the cattle did consume it much more readily than in years when other forages are more abundant.  The drought even impeded the germination of the S. lespedeza seed.  The M.S. graduate student that started in August of 2023 went to several locations and collected seed from plants.  She then did a germination test and 0% of the seeds germinated.  These are some interesting observations found from the ranches as even though there was not many plants in the pastures and seed production to the seed bank was limited over 2022 and 2023, there is such a large seedbank that if we have a good moisture year in 2024 there will be heavy stands of this plant, thus continuing the need to try and keep reducing seed production and plant persistence.

Project 3: Will be completed in summer of 2024

Project 4: Due to the extreme drought in the area where the virtual fences were set-up, there was no winter annual forages that grew, therefore we could not graze them.  We have developed the satellite towers into a mobile unit and will place some collars on cows that are on graze-out rye or wheat in the spring of 2024 to test the ability to use virtual fences in cropping systems.

Research conclusions:

We feel confident that virtual fences (the ones evaluated in the year 2023) do not work as a reliable method to restrict cattle to specific water locations. The issue with the current technology in the extremely hilly and hard wooded area where we were located is that there is was a greater than 15 meter (~45 feet) margin of error with location of the animal and the "fence".  For ponds, when including that margin of error we had to have an opening for drinking that was over 60 meters (~180 feet) to guarantee only 15 meters (~45 feet) of water access, consistently.  Unfortunately, the cows we used were so well trained that when they heard the audio sound they immediately backed up and we had some cattle that did not want to try again.  Since water in such a vital nutrient, when we found those cows that were extremely timid and not making it to the water source, we shut the collars off for those animals and they were able to drink anywhere, which defeated the purpose.  Because we deemed a 180 foot (60 meter) area that may or may-not allow cattle access to drink as too large and inconsistent in allowing animals to drink, that this currently is not a viable option.  With improvements in technology and more accurate location of the animal and virtual fences this may become an option, but to-date it is not.  The technology has been useful to keep cattle off of riparian areas, but if you need to use that as a water source, the technology needs to keep improving.

Participation Summary
3 Farmers participating in research

Education

Educational approach:

A graduate student has been working on understanding pasture management practices that include grazing and water management.  She began her Master's program in August 2023.  To date she has identified 5 different methods that have been used to try and manage for the invasive plant, S. lespedeza.  To date, she has been familiarizing herself with best management practices for cattle production and ecosystem management and will begin collecting data in the summer of 2024.

Project Activities

Eastern Kansas Grazing School

Educational & Outreach Activities

4 Consultations
1 On-farm demonstrations
1 Tours
1 Workshop field days

Participation Summary:

12 Farmers participated
6 Ag professionals participated
Education/outreach description:

The Southeast Kansas Grazing School was hosted on the G3 cattle operation in 2023.  As part of the grazing school we measure forage availability, calculate stocking rates, and talk about best management practices.  In addition to those basics, the employees of G3 Cattle company where the virtual fence system is set-up, showed the system to the attendees of the event and were involved in a panel with other grazing experts discussing the good things, and the issues observed with the system utilized.

A tour was given to a few members of the local NRCS and word-of-mouth has resulted in 4 other producers calling the Extension and Research Center for more information about the system.

Learning Outcomes

12 Farmers reported changes in knowledge, attitudes, skills and/or awareness as a result of their participation
6 Agricultural service providers reported changes in knowledge, skills, and/or attitudes as a result of their participation
Key areas taught:
  • Grazing management practices

Project Outcomes

6 Farmers changed or adopted a practice
Key practices changed:
  • Added rotational grazing to their production system

  • Developed a complementary forage base system for operation

  • Improved pasture inventory methods (ID of weeds, forbs, legumes, etc)

Success stories:

A cattle producer from Southeast Kansas has decided to convert 1/3 of his current land that is in tall fescue (cool-season perennial grass) into Bermudagrass (warm-season perennial grass).  This in conjunction with rotational grazing has him estimating being able to increase his cowherd capacity by 10% and should increase number of calves to sale at weaning by 22%.

The owners of the cattle that had the collars found a great success with use of these collars.  Historically, they calve first calf heifers in a small pasture near the headquarters, then move them to a very large summer pasture.  When releasing the first time mothers into this large pasture they would run off and leave their calves and send several days finding their calf.  The producers were able to successfully build a virtual fence in an 1/8 of the pasture and hold the pairs there for a few days to get them matched up.  By building this virtual fence they pairs were matched within hours of turnout which was helpful for the health of calf and maintaining the mothering ability of the dam.

Any opinions, findings, conclusions, or recommendations expressed in this publication are those of the author(s) and do not necessarily reflect the view of the U.S. Department of Agriculture or SARE.