Conservation Evidence strives to be as useful to conservationists as possible. Please take our survey to help the team improve our resource.

Providing evidence to improve practice

Action: Use wire fencing to exclude large native herbivores Forest Conservation

Key messages

Read our guidance on Key messages before continuing

  • Five of ten studies (including two replicated, randomized, controlled  studies) in Australia, Bhutan, Canada, France, Portugal and the USA found that using wire fencing to exclude large herbivores increased the cover and  size of understory plants. Six studies found no effect of wire fencing on the cover, seed density, species richness and diversity of understory plants.
  • Two of the above studies and one paired-sites study in Ireland examined the effect of using wire fencing to exclude large herbivores on young trees. One found it increased the biomass, one found it decreased the density of young trees and one found mixed effects depending on the species.
  • Two replicated, controlled studies in the USA found that using wire fencing to exclude large herbivores increased tree density.

 

Supporting evidence from individual studies

1 

A replicated, controlled study in 1997 in eucalypt woodlands in New South Wales, Australia (Grant & MacGregor 2001) found no effect of excluding large herbivores on topsoil seed density. The density of topsoil seeds was similar between grazed and ungrazed plots (12,360 and 9,351 seeds/m2 respectively). Topsoil (0-10 cm depth) seeds were monitored in five grazed (1-2 dry sheep equivalents/ha/year) and five ungrazed 20 ×20 m plots. Soil was sampled in 1997 using a soil corer 20 cm in diameter.

 

2 

A replicated, randomized, controlled, study in 1977-1981 in a subtropical moist lowland forest in Alabama, USA (Brockway & Lewis 2003) found that excluding deer and cattle had no effect on plant cover, species richness or diversity after four growing seasons. Plant cover were 135, 132 and 138, numbers of species were 29, 29, 29 and species diversities (Shannon’s index) were 2.37, 2.41 and 2.46 for ungrazed, deer-grazed and cattle-and-deer-grazed treatments respectively. Three 900 ha allotments, each containing six 150 ha blocks were established in 1977. Three treatments were randomly assigned to three 0.1 ha plots within each block: grazing by deer and cattle excluded, grazing by deer only and grazing by deer and cattle. Plant data were collected in September and October 1978–1981 along three 20 m line transects within each treatment plot.

 

3 

A replicated, randomized, controlled study in 1996-2004 in temperate mixed forest in Tennessee, USA (Webster, Jenkins & Rock 2005) found no effect of excluding deer on spring flower species richness and diversity. Numbers of species (exclosure: 1.5-6; unfenced: 2-6/100 m2) and species diversity (Shannon's index exclosure: 0.25-0.75; unfenced: 0.25-0.90) were similar between treatments. Data werecollected in 2004 in five exclosure (fenced to exclude deer browsing in 1996) and five control (unfenced) plots (10 × 10 m) in each of three sites.

 

4 

A replicated, paired sites study in 1969-2001 in temperate broadleaf forest in Ireland (Perrin, Kelly & Mitchell 2006) found that excluding deer decreased the number of seedlings but increased the number of saplings and the height of common holly Ilex aquifolium and rowan Sorbus aucuparia. In yew Taxus baccata wood sites, the density of holly seedlings was lower in fenced plots (fenced: 0.4; unfenced: 2.1/m2) , whereas the density of rowan seedlings was similar between treatments (fenced: 0.2; unfenced: 0.2) Sapling density of both holly (fenced 0.7, unfenced <0.1) and rowan (fenced 0.4, unfenced 0.0, respectively) and juvenile height () ( holly: fenced 45, unfenced 8cm; rowan: fenced 70, unfenced 10 cm) was higher in fenced plots. In oak-wood sites, seedling density for both holly (fenced: 0.5; unfenced: 21.9) and rowan (fenced: <0.1; unfenced: 0.8) was lower in fenced plots. Sapling density for holly was higher in fenced plots (fenced: 3.0; unfenced: 0.5) and for rowan it was similar between treatments (fenced: 0.3; unfenced: <0.1). Sapling juvenile height was higher in fenced plots for both holly (fenced: 130; unfenced: 10) and rowan (fenced: 240; unfenced: 10). Data were collected in 2001 in three fenced plots in yew wood-type sites (764-1,036 m2 deer-proof exclosures established in 1969-1970), four fenced plots in oak wood-type sites (225-1,090 m2, established in 1974-1975) and seven adjacent unfenced plots (225-600 m2).

 

5 

A replicated, paired-sites, before-and-after study in 1997-2005 temperate mixed conifer forest in the Bhutan Himalayas (Darabant et al. 2007) found that excluding large herbivores increased bamboo Yushania microphylla growth but decreased seedling density of all conifer trees, particularly Himalayan hemlock Tsuga dumosa and Sikkim spruce Picea spinulosa. Eight years after treatment, the percentage cover of bamboo increased by 42% in grazed and 58% in ungrazed plots. The number of all conifer tree seedlings increased by 16,333/ha in grazed and only 166/ha in ungrazed plots. The number of Himalayan hemlock seedlings increased by 14,417/ha in grazed and decreased by 167/ha in ungrazed plots. The number of Sikkim spruce seedling increased by 667/ha in grazed and decreased by 166/ha in ungrazed plots. In 1996, five pairs of 4×6 m treatment plots: grazed (unfenced) and ungrazed (fenced to keep out large herbivores) were established in each of two sites. Each was divided into six 2×2 m subplots that were sampled repeatedly in 1997 at the time of treatment and again in 2005.

 

6 

A replicated, paired-sites, before-and-after trial study in 1998-2006 in temperate broadleaf forest in Quebec, Canada (Collard et al. 2010) found that excluding deer  increased the above ground biomass of spring-flowering herbaceous species, small seedlings and large shrubs and trees, but not of summer-flowering herbaceous species, grasses, ferns and small deciduous shrubs. Eight years after treatments, the above ground biomass of small and large spring-flowering herbaceous species had increased by 119% and -19% in grazed plots and 570% and 89% in ungrazed plots respectively. The biomass of small deciduous seedlings had decreased by 63% in grazed and 18% in ungrazed plots. The biomass of large deciduous shrubs and trees had increased by 99% in grazed and 418% in ungrazed plots. Excluding deer did not affect above ground biomass of summer-flowering herbaceous species, grasses, ferns and small deciduous shrubs. Six sites of two 625 m2 treatment plots: grazed (control) and ungrazed (deer exclosure) were established in 1998. Above ground biomass (g/m2) was estimated in 1998 and 2006 in twenty 2 × 0.1 m subplots in each plot.

 

7 

A replicated, controlled study in 2002-2008 in temperate broadleaf forest in Pennsylvania, USA (Huebner et al. 2010) found that excluding herbivores increased fruit production and the cover of some under-canopy species. Six years after treatment the total number of fruit/plot (fenced: 20-430; unfenced: 0-1), relative cover of the palatable herbaceous species painted trillium Trillium undulatum, sessile bellwort Uvularia sessilifolia and Indian cucumber-root Medeola virginiana (fenced: 0-3%; unfenced: <1%), cover of bramble Rubus spp. (fenced: 1%-25%; unfenced: <1%) and the number of tree saplings (fenced: 0-2; unfenced: <1.0/m2) were higher in fenced than unfenced plots. The cover of hay-scented fern Dennstaedtia punctilobula was similar between treatments (0-70%). Data were collected in 2008 in three blocks of 12 fenced (2 m tall fence with 10 × 10 cm openings) and 12 unfenced plots (50 × 80 m). Plots were established in 2002 in an area subjected to high and constant deer herbivory pressure.

 

8 

A replicated, controlled study in 2005-2008 in temperate forest in France (Pellerin et al. 2010) found no effect of excluding deer browsing on species richness and diversity of trees and herbaceous species. The number of woody plant species (deer exclusion: 8-10; unfenced: 7-10/m2) and their species diversity (Shannon's index deer exclusion: 2.1-2.5; un-fenced: 1.9-2.1) and the number of herbaceous species (exclusion: 17-20; un-fenced: 13-17/m2) and their species diversity (Shannon's index deer exclusion: 3.4-3.5; unfenced: 2.9-3.1) were similar between treatments. Data were collected in May 2008. At one site there were 60 sampling plots (1 m2) inside a 1 ha fenced area (deer exclusion) and 60 similar plots inside a 1 ha open area (unfenced). At a second site there were 42 sampling plots (1 m2) inside a 1.5 ha fenced area (deer exclusion) and 42 similar plots inside a 1.5 ha open area (unfenced). Both sites were regularly grazed by roe deer Capreolus capreolus and red deer Cervus elaphus. Exclosures were set up in March 2005.

 

9 

A replicated, controlled study in 2001-2006 in Mediterranean-type shrubland in California, USA (Potts, Marino & Stephens 2010) found that excluding deer increased shrub height. Shrub height was higher in deer exclusion (68 cm) than in unfenced plots (55 cm). Five unfenced control and five deer exclusion (1.5 m fence constructed in 2001-2003) plots (2.5 m2) were replicated in twenty areas (2 ha). Data were collected three years after treatment.

 

10 

A replicated, paired-sites study in 1979-1990 in Mediterranean oak woodland in south-east Portugal (Bugalho et al. 2011) found that excluding red deer Cervus elaphus and fallow deer Dama dama increased the biomass of herbaceous species and the relative cover of legumes Fabaceae, but did not affect the number of plant species. The biomass of herbaceous species was 177 g/m2 in ungrazed and 100 g/m2 in grazed plots. Relative cover of legumes was 10% in ungrazed and 5% in grazed plots. The total number of plant species was similar in grazed (44) and ungrazed (42) plots. Five blocks of paired ungrazed (fenced) and grazed (unfenced, grazed mainly by red deer and fallow deer) plots (25×25 m) were established in the study area in 2001. In 2003, plant biomass and the relative cover of plants were measured in four subplots (2×4 m) within each plot.

 

11 

A replicated, controlled study in 1981-2010 in Mulga Acacia aneura dry forest in Queensland, Australia (Fensham, Silcock & Dwyer 2011) found no effect of excluding herbivores on the number of plant species. There was no difference between treatments for species richness of all plants (exclusion: 15; unfenced: 16 species/plot), annual grasses (exclusion: 2; unfenced: 3), perennial grasses (exclusion: 3; unfenced: 3), annual herbaceous species (exclusion: 5; unfenced: 5) and perennial herbaceous species (exclusion: 4; unfenced: 3).. In 1981-1983, two treatments (50 × 50 m plots) were replicated at three sites: control (unfenced) and fences to exclude all mammalian herbivores >200 g. Plant species richness was determined in 2008 in twenty 2 × 7 m subplots in each treatment.

 

12 

A replicated, controlled study in 2000-2007 in temperate conifer forest in Oregon, USA (Endress et al. 2012) found that excluding grazing herbivores increased the density of tree species. The combined density of Populus spp. and willows Salix spp. was higher in herbivore exclusion (212 trees/ha) than in unfenced plots (66). The density of the most common species, cottonwood P. trichocarpa was 122 trees/ha in herbivore exclusion and 24 trees/ha in unfenced plots. Two 1 ha plots, one in an area with grazing by cattle Bos taurus, elk Cervus elaphus, and mule deer Odocoileus hemionus and one fenced herbivore -exclusion area were established in each of six sites. Data were collected from 2005 to 2007.

 

13 

A replicated, controlled study in 1987-2008 in boreal forest in Minnesota, USA (White 2012) found that excluding deer and snowshoe hares Lepus americanus increased tree density, basal area and biomass. Increases were higher in exclusion plots for tree density (unfenced: 81%, 1,617 to 3,219 /ha; exclusion: 274%, 1,375 to 4,836 /ha), basal area (unfenced: 50%, 15 to 23 m2/ha; exclusion: 125%, 11 to 25 m2/ha) and biomass (unfenced: 37%, 72 to 98 tons/ha; exclusion: 95%, 53 to 104 tons/ha). Data werecollected in 1991 and 2008 in three exclusion (fenced to exclude deer and snowshoe hares in 1987-1990) and three control (unfenced) plots (0.25/ha).

 

Referenced papers

Please cite as:

Agra H., Schowanek S., Carmel Y., Smith R.K. & Ne’eman G. (2019) Forest Conservation. Pages 331-347 in: W.J. Sutherland, L.V. Dicks, N. Ockendon, S.O. Petrovan & R.K. Smith (eds) What Works in Conservation 2019. Open Book Publishers, Cambridge, UK.