Every rancher dreads getting that phone call - "Your cows [sheep, goats, etc.] are out." And anyone who relies on fences to keep livestock contained has probably received that call at some point. Fences fail, gates are left open, somebody forgets to hook up the electric fence energizer. Whenever I get that call, I drop everything else and take care of getting our sheep back where they belong. Getting our livestock guardian dogs (LGDs) back in the sheep paddock can sometimes be more problematic!
Over the years, I've come to realize the importance of a well-bonded LGD. Even if our sheep escape, a well-bonded dog will stay with them - and will often come back with the flock when we herd them back to the paddock. But sometimes, a dog will get out of our fences to chase off a predator. Sometimes a dog will simply decide to explore the neighborhood. Sometimes a dog will slip a collar or scoot through a gate when we're moving the sheep.
LGDs can get into trouble when this happens. One of our earlier dogs, Reno, loved chicken dinners - woe to the free-range chickens that might be nearby. He also disliked outdoor cats immensely (and often to their detriment). And so on the occasions that he got out of the sheep paddock, I was often in a hurry to catch him and keep him out of trouble. I'd call to him and follow him around, trying to catch him by the collar. The more I called (often increasingly frantically), the more he'd run away from me. I joked that if he'd had five toes on a front paw, he would have flipped me off!
I discovered, however, almost by accident, that he would generally come back if I ignored him. One afternoon, he escaped and took off across the ranch. I went about fixing fence and checking sheep, and within five minutes, he was back and wanted to be back with his sheep. I've subsequently experienced the same thing with other dogs.
This morning, I got that call - "Your sheep are out." When I arrived, I found most of the sheep grazing in a neighbor's pasture - and spotted Dillon the LGD gallivanting across the far side of the pasture, perhaps a quarter-mile away. I focused on getting the sheep back into their paddock, and before I finished Dillon returned and allowed himself to be herded along with the sheep. A few minutes later, I found several straggler sheep outside another section of fence. As my border collie brought them back, I opened the electronet for them - allowing Dillon to escape again. Once again, I ignored him - and within minutes he walked up to me so I could catch his collar.
This kind of behavior, I think, is related to the bonding process that we use. I want my LGDs to know how to ride in the truck, to accept being walked on a leash or tied out on a chain while we're working sheep. But I most want them to want to be with their sheep in all circumstances. Teaching a livestock guardian dog to come when I call (or other obedience training, for that matter) seems to require a bond with me rather than with the livestock. A dog that sits, stays, and comes when I call, might prefer to be with me rather than with my sheep. Working LGDs are not pets, and so we have to meet them on their terms when we need them to guard livestock. Figuring this out has made catching the occasional wayward dog much less stressful!
As I've written numerous times, research into the efficacy of livestock protection tools, including livestock guardian dogs, is difficult (if not impossible). The traditional model for scientific inquiry - that of comparing a treatment to a control - is extremely challenging when it comes to livestock protection tools. Fundamentally, nobody wants to be part of the control group (that is, nobody wants to leave a group of livestock unprotected to see if the treatment works!). Further challenges arise when we begin thinking about other variables - questions like the specific environment, the predators in that environment, the dietary preferences of those predators, the surrounding land uses, and so on.
Yet these challenges don't mean that we shouldn't try to shed light on questions about where specific livestock protection tools may work, or where they may fail. I like the idea of doing case studies - real world examples of the success or failure of these tools. In my mind, a useful case study would objectively describe as many of the site- and operation-specific details as possible. Case studies could take into account that many real-world management systems employ multiple tools. And case studies could be important whether or not a particular approach successfully prevented predator losses - sometimes we learn more from our failures than from our successes. The following account, then, is my first attempt at writing one of these case studies.
Flying Mule Sheep Company grazes approximately 100 head of sheep on foothill annual rangeland west of Auburn, California, from mid-December through early April. The flock is comprised of bred ewes (approximately 80 head) and replacement yearling ewes (approximately 20 head). The grazed landscape is a large-lot subdivision (20-40 acre lots). Individual parcels are connected via paved and unpaved private roads and Nevada Irrigation District canals. Many residences have domestic dogs; some have horses and donkeys. Vegetation in the grazed landscape includes open grasslands, blue/live oak savanna, blue/live oak woodland, and riparian vegetation. Surrounding land uses include grazing land (cattle, sheep, and goats) and a large regional park (mostly wildland).
Twelve game cameras were placed throughout the grazed landscape in late December. Cameras were placed adjacent to game trails, roads, and canals to help determine the species of wildlife present and the frequency of camera "capture" in relationship to the proximity of livestock guardian dogs and sheep. In order of prevalence in game cameras from late December through early April, I noted coyotes, foxes, bobcats, and a single mountain lion (in the evening on March 1, 2020). Other wildlife caught on camera included deer, raccoons, skunks, jackrabbits, and turkeys.
During the graze period (December 15, 2019 through April 6, 2020), we had no predator losses. In early February, I found a buck that was likely killed by a mountain lion. On the night that we documented the mountain lion in a game camera (March 1), the flock was in a 13-acre paddock, the boundary of which was about 30 yards away from the camera location. On that date, there were 47 lambs with the ewes (between the ages of 1 day and 11 days). The sheep had been moved into this paddock on the morning of March 1. We lost three lambs during the time the sheep were in that paddock due to starvation or mis-mothering.
From a purely scientific standpoint, I cannot say that the dogs and electric fence prevented predation. While the cameras clearly demonstrated that we had predators in the vicinity of the sheep, I don't know that these specific predators would have killed sheep (rather than wildlife prey) if they'd had the opportunity. I don't know if these predators took livestock from unprotected herds/flocks during the same time period. That said, I can conclude that I feel much safer having dogs with the sheep in this landscape! I can also conclude that the mountain lion I caught in my camera has probably seem me more than I've seen it!
Like any livestock management tool, livestock guardian dogs come with both costs and (hopefully) benefits. Some of these are easily calculated - for example, through today, we've spent $624.70 on dog food and veterinary costs related to our livestock guardian dogs this year. We currently have 3 dogs (Bodie, a 3-year-old I purchased as a pup for $350; Elko, a 2-year-old given to me as a pup; and Dillon, a 9-month-old pup purchased for $500). Some of the costs and benefits are less easily calculated, however; how do I know how many sheep didn't die because we had dogs with them? What is the value of my own peace of mind? A recent paper by Dr. Ellen Bruno (Cooperative Extension specialist in agriculture and resource economics at UC Berkeley) and Dr. Tina Saitone (CE specialist in agriculture and resource economics at UC Davis) sheds new light on these questions. Read the complete paper here.
Using data from the University of California's Hopland Research and Extension Center, Bruno and Saitone estimated that dogs reduced lambs lost to coyotes by 43% each year; ewe losses were reduced by 25%. The authors calculated the present value of these prevented deaths over the 7-year useful life of the dogs to be $16,200 (present value calculations were based on the market value of the lambs as well as the value of running-age ewes). Their model was based on using one dog per 100 ewes (more on this below).
On the cost side, the authors included initial purchase of pups, dog food (and labor associated with feeding the dogs), veterinary costs, and dog replacement costs. Labor costs, as they note, are largely dependent on the type of production system - Hopland's labor costs are probably much higher than the typical commercial operation. Using net present value analysis, Bruno and Saitone found that the costs of Hopland's livestock guardian dogs exceeded the benefits (in the value of lambs and ewes not killed by predators) by $13,412 over the seven-year analysis period. In other words, the dogs didn't pay their own way.
Bruno and Saitone offer several important caveats when interpreting these results. First, many ranchers report that dogs eliminate predation entirely (which has been our own experience). If this had been the case at Hopland, the benefits would have exceeded the costs of using dogs by over $12,000. Second, labor-related expenses associated with dogs can be difficult to quantify. In our operation, feeding the dogs is part of our daily check of fences and sheep - we see the sheep every day whether we have dogs with them or not. We charge about 5 minutes per day to feeding 3 dogs - even if I pay myself $20 per hour for this work, our "dog" labor amounts to $371 per dog annually. Hopland, on the other hand, reported labor costs of nearly $1,600 per dog per year. Finally, the authors note that lamb and ewe prices may (and usually do) change from one year to the next - and sometimes dramatically. Sheep values can alter the cost:benefit ratio.
Skeptics might wonder, "Even if you use dogs, if you're not experiencing any predator losses, maybe there aren't any predators around." My ongoing research into livestock guardian dog behavior suggests that there are ALWAYS predators around where small ruminants are grazing (whether on rangeland or irrigated pasture). Using trail cameras, we frequently "capture" coyotes, foxes, and bobcats within 10-15 feet of our sheep paddocks. Interviews with sheep- and goatherders working in the Sierra Nevada indicate that coyotes are heard - and often seen - every night near sheep and goat bed grounds. Though we see them less frequently, we know there are mountain lions and black bears in the vicinity of these operations. The predators are there - the dogs must be at least partly responsible for the lack of predator losses!
As I've written previously, the number of dogs used by producers can vary greatly - from one producer to the next, and from one season to the next on the same operation (see How Many Dogs?). One of the bands of sheep I'm observing near Truckee is guarded by a single dog (band is roughly 1,000 ewes - this scenario is significantly more cost effective than the 1 dog per 100 ewes ratio used in Bruno and Saitone's model). This ratio works because the band is comprised of mature ewes without lambs - and because the predators have plenty of other prey at this time of year. Once this band moves back to Los Banos to lamb on alfalfa stubble later this fall, the dog-to-sheep ratio will increase.
In my experience, peace of mind for the shepherd (or goatherd) can be a significant (if qualitative) benefit. My friends Brad Fowler and Nathan Medlar recently started a targeted grazing project at Squaw Valley Ski Resort north of Lake Tahoe (see Watching Other Dogs). They started the project without livestock guardian dogs (mostly to avoid conflicts with recreationists). They are herding the goats on the ski slopes during the day and penning them at night near their camp (a tent on the side of the mountain). Brad reported that neither they nor the goats slept at all on the first night - the coyotes kept the goats stirred up even though they were protected by electric fence. Brad and Nathan added two dogs on the second day - which relaxed the goats (and the goatherds). Brad reported both herders and livestock slept soundly on the second night.
Finally, research at the U.S. Sheep Center in Dubois, Idaho, found "that ewes grazing with accompanying LGD will travel greater daily distances compared with ewes grazing without LGD accompaniment. As a result of traveling greater distances, ewes may also be exposed to more and varied foraging opportunities." See Webber et al. 2015 for the complete study. To me, this suggests that dogs may make our grazing operations more efficient - allowing us to access forage that would otherwise not get grazed by unprotected livestock. This increased grazing efficiency can reduce our supplemental feed costs.
From my perspective, perhaps the most important part of Bruno and Saitone's work comes at the end of the paper:
"Sheep producers who are considering the purchase of LGDs, or those who already have LGDs and are interested in their return on investment, need a few pieces of data to make this determination. Market lamb and ewe prices are typically well known to producers and can be used, in conjunction with efficacy rates from this study, to estimate the benefits of LGDs.
"On the cost side, producers would need to make some logical forecasts about the time required to maintain LGDs, given their operation specifics.... Also, using guidance from the literature included herein, producers could calculate the likely dog cull and mortality costs of the LGD's useful life."
Ultimately, the success of any livestock protection tool (including lethal control) is highly variable depending on operator characteristics and environmental conditions. Dogs work in our operation because we see the sheep every day and because they are our only option for protecting lambing ewes (we lamb on pasture without access to a lambing shed). Dogs work for the range outfit on the Tahoe National Forest as well; human presence, the vigilance of the dogs, and the stage of production during their time in the mountains virtually eliminates predator losses. And dogs work for the targeted grazing outfits I work with in the foothills and mountains; peace of mind and lack of predator losses justify the costs of keeping dogs in these operations, too.
As some readers of this blog may know, I'm currently working on a research project examining livestock guardian dog behavior. The back story is this: several years ago, I was invited to demonstrate electro-net and livestock guardian dogs at a workshop on livestock protection tools. The electro-net fencing was easy! However, since I was speaking at midday, the LGD demo was less than dynamic - the dog came over to the fence, barked half-heartedly at the people he didn't recognize, and resumed napping in the shade!
This experience got me thinking! How could I demonstrate the effectiveness of these dogs without dragging folks out to observe the sheep in the middle of the night (when the dogs are much more active)? Geographic positioning system (GPS) technology seemed like a possible answer - but commercial GPS collars were too expensive for my cooperative extension / sheepherder budget. While perusing Facebook one day, I ran across a post from Dr. Derrick Bailey at New Mexico State University. Dr. Bailey was using home-built GPS collars to track cattle distribution on New Mexico rangeland! At last, an affordable solution! Dr. Bailey was gracious enough to spend an hour on the phone with me talking about my project ideas - and he shared the technical details of the collars he was using.
Here's a quick photo guide to building the collars I'm using on LGDs (and on sheep). The materials include:
- LGD collars from Premier 1 Supplies (I like these extra-wide collars - I think they're comfortable for the dogs, and they seem to hold up in rangeland conditions). https://www.premier1supplies.com/p/guard-dog-collars?cat_id=164
- 3-1/2" x 2" threaded nipples and threaded caps (for the case)
- 1/2" x 5/32" pop rivets and #8 SAE flat washers (to attach the case to the collar)
- i-gotU GT-600 travel and sports logger (available on Amazon - https://www.amazon.com/i-gotU-USB-Travel-Sports-Logger/dp/B0035VESMC/ref=sr_1_2?crid=2OO21VMYVPBN2&keywords=i+got+u+tracker&qid=1564451973&s=gateway&sprefix=I+got+U%2Caps%2C194&sr=8-2
The collars take about 5 minutes to build. The i-gotU trackers can be programmed to collect GPS coordinates from every 5 seconds up to every 5 minutes. Set at 5 minute intervals, the batteries in the unit will last 10 days. Dr. Bailey also sent me plans for an auxiliary battery system - that will be my next project!
I've also experimented with an Optimus 2.0 tracker (https://www.amazon.com/Optimus-Tracker-6543857646-GPS-2-0/dp/B01C31X50K/ref=sr_1_4?crid=UR8F2VQBBT8M&keywords=optimus+tracker&qid=1564452200&s=gateway&sprefix=optimus+t%2Caps%2C199&sr=8-4) which sends a real-time signal to my cell phone with the position and speed of travel of the unit. These trackers don't record positions, but they are useful from a practical standpoint - they will send an alarm to my phone if a guard dog is out of my pasture.
I'm hoping that we'll have some data to share from my project on the Tahoe National Forest north of Truckee in the next couple of weeks. Working with Talbott Sheep Company, I've collared 2 dogs in each of 2 bands of sheep. So far, the collars seem to be working great!
And on a humorous note, as you can see from the photos, I put UCCE (for University of California Cooperative Extension), along with my phone number, on the collars. I received a text yesterday that said:
"Hello, we found Ucce at the upper little truckee campground this morning. He still has his tracker around his neck and is just hanging out at the campsites."
I explained that we were doing a research project with the dogs and that someone would come by to get the dog soon.
That said, I think Ewecie (or maybe Ewechie) would be a great name for a guard dog, don't you!?
Here are some photos to walk you through building a collar.
One of the questions I'm asked most frequently when it comes to livestock guardian dogs is, "How many dogs do I need to protect my sheep/goats/cows?" As you might imagine, the short answer is, "It depends." The long answer is more complex. From an economic perspective, the answer is, "As many as it takes to hold predator losses in your operation at an acceptable level, but no more than that." From a production perspective, I've found that the answer depends on operational characteristics, the environment, and the abilities of the specific dog(s).
While it is tempting to try to develop a rule of thumb recommendation (like one dog per 100 ewes), reality is usually more complicated. Wearing my sheepherder economics hat for a moment, the fundamental question comes down to comparing the costs of a dog versus the benefits the dog provides. On the cost side of the ledger, I must account for the cost of dog food, veterinary care, and depreciation. In our operation, these annual expenses amount to roughly $600 per dog. On the benefit side of the ledger, I need to know how many sheep DON'T get killed by predators to determine if my $600 in expenses are justified. Obviously, this is not an easy number to estimate - how can I measure something that doesn't happen? How do I quantify the sheep that might have died had I not had a livestock guardian dog with them? I suspect we'd lose more sheep if we didn't use dogs, but I'm not willing to leave the sheep unprotected to find out!
Operational characteristics, in my experience, play a significant role in determining the optimal number of dogs. Birthing seasons (spring vs. fall), other livestock protection tools (like electric fence, on-site herders, night penning, etc.), grazing management (set stocking versus rotational grazing), and the number of individual herds or flocks all factor into determining the right number. Using our operation as an example:
- We lamb in the late winter and early spring, when there is not a significant natural prey base for the wild predators in our environment. Our lambing paddocks are 7 miles from our home. This argues for more dogs.
- We use electro-net fencing, which definitely deters canine predators (dogs, coyotes and foxes) as well as bobcats. This allows us to get by with fewer dogs.
- We move the sheep frequently - they move to fresh pasture every few days, and graze different properties in spring/summer versus fall/winter. I suspect all of this movement keeps the predators off balance. This allows us to get by with fewer dogs.
- We rarely (if ever) have all of our sheep in one mob. This time of year, the mature ewes are in one flock; the feeder lambs and replacement ewe lambs are in a second flock; the rams in a third location. During breeding season, we have two separate breeding groups plus a group of lambs. This argues for more dogs.
Based on these factors, we feel that we need at least three dogs for our small, part-time operation. With three dogs, we can protect three different groups of sheep or place two dogs together during our most vulnerable time of year (lambing). During some parts of the year, we have more dogs than necessary, which provides flexibility if we begin to have problems with predation.
The environment where we're grazing, and the predators it contains, is a second critical consideration. Here in the Sierra foothills (at least at the moment) our main predators (in order of potential threat) are domestic dogs, coyotes, mountain lions, black bears, bobcats, foxes, and birds of prey. I've spoken with ranchers on the north coast who would add crows, ravens, and magpies to that list. And ranchers in northeastern California would add gray wolves. Predator density and prey base also come in to play. Are there several established packs of wolves in the region? Is there sufficient native prey? Are these particular wolves (or coyotes, or mountain lions) known to prey on livestock? Each of these questions are important to consider when determining how many dogs a particular operation might need.
Finally, every livestock guardian dog is an individual. Some are athletic and want to patrol a wide area; others want to stay with their livestock. Some dogs are more canine aggressive than others (an important trait in wolf habitat); others will harass bears. And these traits will change over time - a dog that was aggressive and athletic in his younger days might be content to stay with lambs on irrigated pasture in his later years. In my experience, there is more variation between individuals than there is between livestock guardian dog breeds (a subject for a future blog post!).
Finally, I started a new phase of my livestock guardian dog behavior study this week. I'll be tracking the movements of four dogs (2 each in separate 1000-ewe bands of sheep) grazing on the Tahoe National Forest in Nevada and Sierra Counties (in an area that a collared Oregon wolf has been known to visit in the last 12 months). This is a long-time producer with experienced herders operating on open range with no fences. They typically use two dogs with one band and three dogs with the other, and experience less than one percent death loss while the sheep are on Forest Service allotments. They also have additional dogs they can add to each band if predator problems begin to escalate.
I think this illuminates the "it depends" answer in my first paragraph! They have 1 dog per 400 sheep; we have 1 dog per 51 sheep. They are grazing mature ewes in a relatively wild environment for only 75 days - and at a time when the natural prey base is plentiful. We need more dogs to protect ewes and lambs at an especially vulnerable time of year (and I should note - the large operation needs more dogs at lambing as well). The common thread for each of these operations, however, is that we are constantly evaluating our need for predator protection against the cost of providing it. If we could get by with fewer dogs, we would; similarly, if the large operation needs more dogs this summer, they'll add dogs. In other words, it depends!