Use guardian animals (e.g. dogs, llamas, donkeys) bonded to livestock to deter predators to reduce human-wildlife conflict

How is the evidence assessed?
  • Effectiveness
  • Certainty
  • Harms

Study locations

Key messages

  • Twelve studies evaluated the effects of using guardian animals (e.g. dogs, llamas, donkeys) bonded to livestock to deter mammals from predating these livestock to reduce human-wildlife conflict. Four studies were in the USA, two were in Kenya and one each was in Solvakia, Argentina, Australia, Cameroon, South Africa, and Namibia.





  • Human-wildlife conflict (12 studies): Four of seven studies, (including four site comparison studies), in the USA, Kenya, Solvakia, Australia and Cameroon, found that guardian animals reduced attacks on livestock by predators. The other three studies reported mixed results with reductions in attacks on some but not all age groups or livestock species and reductions for nomadic but not resident pastoralists. Two studies, (including one site comparison study and one before-and-after study), in Argentina and Namibia, found that using dogs to guard livestock reduced the killing of predators by farmers but the number of black-backed jackals killed by farmers and dogs combined increased. A replicated, controlled study in the USA found that fewer sheep guarded by llamas were predated by carnivores in one of two summers whilst a replicated, before-and-after study in South Africa found that using dogs or alpacas to guard livestock reduced attacks by predators. A randomized, replicated, controlled study in USA found that dogs bonded with livestock reduced contact between white-tailed deer and domestic cattle.

About key messages

Key messages provide a descriptive index to studies we have found that test this intervention.

Studies are not directly comparable or of equal value. When making decisions based on this evidence, you should consider factors such as study size, study design, reported metrics and relevance of the study to your situation, rather than simply counting the number of studies that support a particular interpretation.

Supporting evidence from individual studies

  1. A replicated study in 1981 of 36 ranches in North Dakota, USA (Pfeifer & Goos 1982) found that guard dogs Canis lupus familiaris reduced sheep losses to predation by coyotes Canis latrans. The average annual predation rate after commencing use of guard dogs (0.4% of the sheep flock) was lower than that before guard-dog use commenced (6%). In 1981, thirty-six ranchers were interviewed about livestock management and losses to predation in the 1976–1981 period. Between them, ranchers had 52 great Pyrenees dogs (44 working and eight training) and two working komondor dogs. All ranchers commenced using guardian dogs during the period. Guarded pastures were 4–486 ha in extent and guarded sheep flocks contained 10–1,300 animals. Dogs were raised with the sheep flock and remained with them most of the time.

    Study and other actions tested
  2. A replicated, site comparison study in 1986 of 134 sheep producers in Colorado, USA (Andelt 1992) found that using livestock-guarding dogs Canis lupus familiaris reduced coyote Canis latrans predation of lambs in fenced pastures and some open ranges, but predation of ewes was not reduced in either. A lower percentage of lambs was killed by coyotes in fenced pastures with livestock-guarding dogs (0%) than without dogs (2–5%). In open ranges, a lower percentage of lambs was killed compared to 20 of 25 producers without dogs (with dogs: 1.2%; without dogs: 16%), this was not the case compared to the five producers without dogs that responded by telephone rather than post (without dogs: 3%). The percentage of ewes killed by coyotes did not differ significantly with dogs (fenced pastures: 0%; open ranges: 0.4%) or without dogs (fences pastures: 0.5–1%; open ranges: 1.1–1.5%). Sheep producers kept ewes and lambs with or without livestock-guarding dogs in fenced pastures (with dogs: 6–7 producers; without dogs: 87–92 producers) or open ranges (with dogs: 10 producers; without dogs: 25 producers). Average flock sizes were 90–321 lambs or ewes in fenced pastures and 910–2,440 lambs or ewes in open ranges. Seven breeds (or mixed breeds) of livestock-guarding dog were used (see original paper for details). The 134 sheep producers responded to postal or telephone surveys in 1986.

    Study and other actions tested
  3. A replicated, controlled study in 1996–1997 on pasture in Utah, USA (Meadows & Knowlton 2000) found that using llamas Lama glama to guard sheep flocks reduced canine predation on lambs in one of two summers. Sheep flocks guarded by a llama lost a lower proportion of lambs to predators in the first summer season than did flocks without llamas. There was no significant difference in losses during the second summer season. Actual loss rates were not presented. Predation rates of ewes and predation in the winter season were very low across all flocks. Coyotes Canis latrans, domestic dogs Canis lupus familiaris and red foxes Vulpes vulpes accounted for 92% of losses to predators. Flocks with llamas averaged 301 sheep (including lambs). Flocks without llamas averaged 333 sheep and lambs. Twenty flocks were each guarded by a single llama. The number of flocks without llamas varied through the study, due to splitting and merging of flocks, from 8 to 29. Sheep producers reported fortnightly, from May 1996 to December 1997, on predation events and flock sizes.

    Study and other actions tested
  4. A replicated, site comparison study in 1999–2000 of savanna across 10 ranches in Laikipia District, Kenya (Ogada et al. 2003) found that at bomas with domestic dogs Canis lupus familiaris in attendance, fewer cattle were killed by predators, though there was no effect on predation of sheep or goats. Fewer cattle were killed by lions Panthera leo, leopards Panthera pardus and hyenas Crocuta crocuta and Hyaena hyaena combined when dogs were present at bomas (0.03 cattle/month) than at bomas without dogs (0.28 cattle/month). There was no significant relationship between dog presence and predation on sheep or goats (data not presented). Livestock were housed in bomas overnight, when 75% of recorded kills occurred. Data on livestock predation and predator deterrence activities at 84 bomas on 10 ranches (nine commercial ranches, one community area) were gathered from ranch managers. Ranches were monitored for 2–17 months, between January 1999 and May 2000.

    Study and other actions tested
  5. A study in 2001–2005 of bushland and savanna in Laikipia and neighbouring districts of Kenya (Woodroffe et al. 2007) found that when livestock were accompanied by one or more domestic dogs Canis lupus familiaris, fewer were attacked by carnivores. Livestock herds grazing by day and those held overnight in thornbrush bomas were less likely to be attacked by carnivores if accompanied by domestic dogs (results presented as odds ratios). Of 502 grazing herds, 24% were accompanied by one or more dogs (average 1.3 dogs/accompanied herd). Of 491 bomas, dogs were present at 71% (average 2.0 dogs/boma). The study documented 105 attacks by spotted hyenas Crocuta crocuta, 96 by leopards Panthera pardus, 44 by African wild dogs Lycaon pictus, 35 by lions Panthera leo and 19 by cheetahs Acinonyx jubatus. From January 2001 to June 2005, eighteen local staff verified reports of livestock lost to predation and gathered data on animal husbandry practices used. Attacked herds or bomas were compared to nearby herds (median 656 m away) or bomas (median 323 m away) that had not been attacked.

    Study and other actions tested
  6. A randomized, replicated, controlled study in 2003 at two forest sites in Michigan, USA (Vercauteren et al. 2008) found that dogs Canis lupus familiaris bonded with livestock reduced levels of contact (and potential for disease transmission) between white-tailed deer Odocoileus virginianus and domestic cattle. In dog-guarded pastures, deer came within 5 m of cattle fewer times (three instances) than in non-guarded pastures (79 instances). No deer were within 5 m of cattle when dogs were present, while 114 events occurred with dogs absent. Deer consumed hay less frequently in dog-guarded pastures (two instances) compared to pastures without dogs (303 instances). At each site, four 1.2-ha pastures, >200 m apart, were enclosed by electric fencing. Deer were baited into pastures with corn and alfalfa. Each pasture contained four calves while two pastures at each site also had a dog. Livestock guarding dogs were great Pyrenees, raised from eight week-old pups, following standard training procedures. Visits of deer into pastures were monitored by direct observation and video surveillance, in March–August 2003.

    Study and other actions tested
  7. A replicated, site comparison study in 2002 on 58 farms in Solvakia (Rigg et al. 2011) found that farms using livestock-guarding dogs Canis lupus familiaris lost fewer livestock to predation than did farms without dogs. The number of livestock lost to predators (mainly grey wolf Canis lupus) in flocks with livestock-guarding dogs (1.1 sheep/flock) was not significantly different to that in unguarded flocks (3.3 sheep/flock). However, dog placement was prioritised at flocks with previously high predation rates. On farms where predation occurred, fewer livestock were lost in guarded (1.5 sheep/flock) than in unguarded flocks (5.0 sheep/flock). Pups (Slovenský čuvač and Caucasian shepherd dog) were reared alongside livestock. Of 34 pups placed on farms in 2000–2004, seventeen were successfully integrated into livestock flocks during the first full grazing season. Reported losses for 2002 were compared between 13 flocks with successfully integrated 1–2-year-old livestock-guarding dogs and 45 farms in the same regions without dogs.

    Study and other actions tested
  8. A replicated, site comparison study in 2005–2011 on a grass-shrub steppe area in Patagonia, Argentina (González et al. 2012) found that use of dogs Canis lupus familiaris by goat herders to guard livestock reduced the killing of predators by herders. Results were not tested for statistical significance. Six of eight herders with working guard dogs reported that they no longer killed predators, one had never done so and one did so less frequently than previously. Nine herders who did not have working dogs all continued to kill predators. Most reported predation was by cougar Puma concolor and culpeo fox Lycalopex culpaeus. Thirty-seven puppies were placed with herders, of which 11 became successful livestock guarding dogs. Herders were interviewed monthly or bimonthly during the dog training period. Nine neighbouring herders without dogs were also interviewed. Interviews included questions about predator control activities carried out by the herders.

    Study and other actions tested
  9. A before-and-after study in 1997–2010 on a grassland-dominated ranch in Queensland, Australia (Van Bommel & Johnson 2012) found that when guardian dogs Canis lupus familiaris were used to protect livestock from dingoes Canis dingo and other predators, sheep mortality declined. By three years after the guardian dog programme commenced, annual sheep losses had fallen to 4% of the flock and remained at 4–7% over the following five years. In the six years before the programme commenced, there was 7–15% annual mortality of the sheep flock. Sheep mortality figures included all causes of death, not only predation. The study was conducted on a 47,000-ha ranch, hosting approximately 12,000–22,000 sheep and 4,000 cattle. Dingoes and feral dogs were the main livestock predators in the area. In 2002, twenty-four Maremma sheepdogs were integrated with the sheep. The sheepdogs worked unsupervised in groups of 1–4. They had access to self-feeders with dry dog food. Dingoes and wild dogs were also baited with poison and wild dogs were shot opportunistically.

    Study and other actions tested
  10. A site comparison study in 2008 of savanna around a national park in Cameroon (Tumenta et al. 2013) found that using dogs Canis lupus familiaris to guard livestock reduced losses through predation among nomadic pastoralists but not among resident pastoralists. Among nomadic pastoralists that owned dogs (53% of all nomadic pastoralists), fewer livestock were lost to carnivores (six animals/year) than among those that did not own dogs (10 animals/year). Among resident pastoralists that owned dogs (33% of all resident pastoralists), there was no significant difference in the number lost to predators (five animals/year) compared to those that did not own dogs (four animals/year). Two hundred and seven resident pastoralists and 174 nomadic pastoralists were interviewed. Subjects reported the incidence of predation on livestock by large carnivores and details of animal husbandry techniques used. Villages were selected based on the tracking of movements of radio-collared lions.

    Study and other actions tested
  11. A replicated, before-and-after study in 2007–2009 of four livestock farms in savanna and shrubland in Eastern Cape, South Africa (McManus et al. 2015) found that using dogs Canis lupus familiaris and alpacas Vicugna pacos to guard livestock reduced attacks by carnivores on livestock, compared to using lethal control of predators. Results were not tested for statistical significance. When guard animals were used, 0–15% of livestock were killed each year by predators, but when lethal predator-control methods were used 5–45% of livestock were killed. Costs of using non-lethal control were lower (0.73–6.02 USD/livestock animal) than were those of lethal control (0.95–­7.94 USD/livestock animal). In August 2006–August 2007, all four farms used lethal methods, including trapping and shooting, to control black-backed jackals Canis mesomelas, caracals Caracal caracal and leopards Panthera pardus. In September 2007–September 2009, farms either used guard dogs (three farms) or alpacas (one farm) to protect animals. Farmers reported the number of livestock killed by predators and associated costs, each September, in 2007–2009.

    Study and other actions tested
  12. A before-and-after study in 2009–2010 of 73 livestock farms in Namibia (Potgieter et al. 2016) found that placing dogs Canis lupus familiaris with farmers to guard livestock reduced the overall number of farmers that killed predators, but increased the numbers of black-backed jackals Canis mesomelas killed by farmers and dogs combined. Eighteen percent of farmers killed livestock predators in the year after dog placement compared to 31% in the previous year. The reduction was larger among subsistence farmers (0% after dog placement; 30% before) than commercial farmers (26% after dog placement; 32% before). However, the number of black-backed jackals killed by farmers and dogs combined in the year following dog placement (3.4/farm) was greater than the number killed by farmers alone the previous year (1.7/farm). There were no significant differences for killings of caracal Caracal caracal (farmer and dog: 0.19; farmer: 0.10), cheetah Acinonyx jubatus (farmer and dog: 0.02; farmer: 0.11) or leopard Panthera pardus (farmer and dog: 0; farmer: 0.02). Anatolian shepherd dogs were placed on 53 commercial farms and 20 subsistence farms. Farmers were interviewed between March 2009 and September 2010. Dogs were placed with a livestock flock at eight weeks old and averaged 39 months old at time of the study.

    Study and other actions tested
Please cite as:

Littlewood, N.A., Rocha, R., Smith, R.K., Martin, P.A., Lockhart, S.L., Schoonover, R.F., Wilman, E., Bladon, A.J., Sainsbury, K.A., Pimm S. and Sutherland, W.J. (2020) Terrestrial Mammal Conservation: Global Evidence for the Effects of Interventions for terrestrial mammals excluding bats and primates. Synopses of Conservation Evidence Series. University of Cambridge, Cambridge, UK.

Where has this evidence come from?

List of journals searched by synopsis

All the journals searched for all synopses

Terrestrial Mammal Conservation

This Action forms part of the Action Synopsis:

Terrestrial Mammal Conservation
Terrestrial Mammal Conservation

Terrestrial Mammal Conservation - Published 2020

Terrestrial Mammal Conservation

What Works 2021 cover

What Works in Conservation

What Works in Conservation provides expert assessments of the effectiveness of actions, based on summarised evidence, in synopses. Subjects covered so far include amphibians, birds, mammals, forests, peatland and control of freshwater invasive species. More are in progress.

More about What Works in Conservation

Download free PDF or purchase
The Conservation Evidence Journal

The Conservation Evidence Journal

An online, free to publish in, open-access journal publishing results from research and projects that test the effectiveness of conservation actions.

Read the latest volume: Volume 21

Go to the CE Journal

Discover more on our blog

Our blog contains the latest news and updates from the Conservation Evidence team, the Conservation Evidence Journal, and our global partners in evidence-based conservation.

Who uses Conservation Evidence?

Meet some of the evidence champions

Endangered Landscape ProgrammeRed List Champion - Arc Kent Wildlife Trust The Rufford Foundation Save the Frogs - Ghana Mauritian Wildlife Supporting Conservation Leaders
Sustainability Dashboard National Biodiversity Network Frog Life The international journey of Conservation - Oryx Cool Farm Alliance UNEP AWFA Bat Conservation InternationalPeople trust for endangered species Vincet Wildlife Trust