Providing evidence to improve practice

Action: Exclude livestock from semi-natural habitat (including woodland) Farmland Conservation

Key messages

  • Seven studies (including four replicated controlled trials of which one also randomized, and a review) from Ireland, Poland and the UK looked at the effects of excluding livestock from semi-natural habitats. Three studies (including one replicated controlled and randomized study) from Ireland and the UK found that excluding livestock benefited plants and invertebrates.
  • Three studies (one replicated controlled and one replicated paired sites comparison) from Ireland and the UK found that excluding grazing did not benefit plants or birds.
  • Two studies (one replicated and controlled, one review) from Poland and the UK found that the impact of excluding grazing as a tool in habitat restoration was neutral or mixed.



Supporting evidence from individual studies


A 1984 review of studies in the UK (Ball 1984) concluded that natural restoration of target heather Calluna vulgaris moorland by removing grazing from upland grassland will happen very slowly. The review describes two long-term studies (1950s-1970s) that monitored botanical changes following exclusion of sheep from upland grassland plots in England (Welch & Rawes 1964, Rawes 1981, 1983) and Wales (Hughes et al 1975, Hill 1983). By 1983, early vegetation changes were slow, mainly involving an altered balance of plant species already present on plots, and entry of heath species was limited. This may have been due to a lack of local seed sources or because seeds were unable to germinate in the close grass turf.

Additional references:

Welch D. & Rawes M. (1964) The early effects of excluding sheep from high-level grasslands in the northern Pennines. Journal of Applied Ecology, 1, 281-300.

Hughes R.E., Dale J., Lutman J. & Thomson A.G. (1975) Effects of grazing on upland vegetation in Snowdonia. Annual Report of Institute of Terrestrial Ecology, 1974, 46-50.

Rawes M. (1981) Further results of excluding sheep from high level grasslands in the northern Pennines. Journal of Ecology, 69, 651-669.

Hill M.O. (1983) Effects of grazing in Snowdonia. Annual Report of Institute of Terrestrial Ecology, 1982, 31-32.

Rawes M. (1983) Changes in two high altitude blanket bogs after the cessation of sheep grazing. Journal of Ecology, 71, 219-235.


A randomized, replicated, controlled trial from 1990 to 1993 on grassland in Northern Ireland (Bell et al. 1994) found that protecting pasture field margins from grazing is likely to improve their wildlife value. When field margins were grazed by sheep (and fertilized), the plant and ground beetle (Carabidae) communities were more similar to those of the open field than to field margins. There were also greater soil temperature fluctuations in grazed margins and their hedges had more gaps. Three treatments and an unmanaged control were replicated three times in the margins of pasture fields on either side of mature hawthorn Crataegus monogyna hedges: fertilized and grazed; ploughed and sown with a game cover strip; and ploughed and left to recolonize naturally. Plants were sampled in July 1991 and August 1993 in quadrats positioned at intervals up to 9 m into each field. Ground beetles were sampled in March, May, July and September each year using three pitfall traps placed 1-2 m and 8-10 m on either side of the hedge.



A replicated, controlled trial in summer 1995 and 1996 in grassland in West Sussex, England (Wakeham-Dawson et al. 1998) found significantly more invertebrates and invertebrate taxa in 1.5 x 0.5 m cages from which grazing sheep were excluded, than in uncaged areas of sheep-grazed grassland (28-38 individuals and 9-11 invertebrate taxa in cages vs 7-15 individuals and 3-5 invertebrate taxa outside). Grass within the grazing-exclusion cages was approximately 40 cm tall, whilst outside the cages it was 2 cm tall. Twelve 10 x 4 m plots, each with one grazing-exclusion cage were established in 1995, with an additional two plots used in 1996. Invertebrates were sampled within and outside the grazing-exclusion cage using a D-Vac suction sampler.



A replicated paired sites comparison in 1999 of grassland habitats on 30 farms in Counties Laois and Offaly, Republic of Ireland (Feehan et al. 2002) (same study as (Feehan et al. 2005)) found that fencing of field and watercourse margins to exclude grazing did not benefit plant diversity. For watercourse margins (eight paired replicates) more plant species were found in unfenced than fenced margins (52 and 56 species on unfenced margins on Rural Environment Protection Scheme and non-Rural Environment Protection Scheme farms respectively, 50 and 48 species on fenced margins on Rural Environment Protection Scheme and non-Rural Environment Protection Scheme farms respectively). For field margins, fencing helped gappy hedges to re-establish and slightly increased the number of ground beetle species in margins. However, as the distance from the hedge to the fence increased (and the ungrazed margin became wider) the number of field margin plant species decreased. The authors recommend that fences to protect hedges should allow stock to graze underneath to maintain the field margin flora. Plants were surveyed in two hedgerows, their associated field margins and one watercourse margin on each farm. Ground beetles were surveyed in June and August.



A replicated paired sites comparison study in 2000 in counties Laois and Offaly, Ireland (Feehan et al. 2005) (same study as (Feehan et al. 2002)) found that fenced watercourse margins on Rural Environment Protection Scheme farms did not have higher numbers of plant species than unfenced watercourse margins on non-Rural Environment Protection Scheme farms (14.7 and 16.1 plant species/margin respectively). Fifteen farms with Rural Environment Protection Scheme agreements at least four years old were paired with 15 similar farms without agreements. On each farm, a randomly selected watercourse margin was surveyed for plants: all plant species were recorded in two 5 x 3 m quadrats, and percentage cover estimated in a 1 x 3 m quadrat within each margin. Eleven of the farm pairs enabled a fenced/unfenced comparison.




A replicated, controlled (paired) study of wet pasture in Leicestershire, UK (Defra 2007) found that bird visit rates were significantly higher in areas with livestock (wet plots: 0.26, dry: 0.10) than in those where livestock had been excluded (wet: 0.17, dry: 0.06). Sampling involved 45 minute bird observations between April 2005-March 2007 (twice/month April-October, once/month November-March).



A series of trials at four upland grassland or heath sites in west Perthshire, Scotland (Holland et al. 2008) showed that excluding sheep caused changes in vegetation and the decline of scarce plant species at two of the sites. Livestock were excluded from three areas of upland grassland and one area of heath in 1998, 2000 or 2001. Vegetation was monitored in one to five years following exclusion. At one chalk grassland site, the species of conservation value: creeping sibbaldia Sibbaldia procumbens, moss campion Silene acaulis and hair sedge Carex capillaris all declined. At the heath site, a population of the small white orchid Pseudorchis albida crashed from 50 spikes in 2001 to none in 2006, after grazing exclusion in 2000. At the two other grassland sites, species composition was not monitored, but the vegetation structure changed, with increased overall height and at one site, less vegetation below 8 cm.


A replicated, controlled study of a degraded species-poor meadow in Central Poland (Klimkowska et al. 2010) found that livestock exclusion had less of an effect on the restoration of plant community composition than topsoil removal and hay transfer. Deep soil removal (40 cm) with hay addition resulted in a community closest to the donor meadows, particularly where grazing was excluded. Species richness in grazed plots was slightly higher after hay transfer (23 vs 18 without transfer); in non-grazed plots hay transfer had no effect. Two plots (35 x 35 m) were subdivided to test combinations of the following treatments: topsoil removal (to 20 or 40 cm), hay transfer from a nearby meadow (collected mid-July 2004-2005, partly dried, stored for 1.5 months, spread in 5-7 cm layer) and livestock/exclusion. Data were obtained from plots on plant species distribution and abundance (2004-2007) and biomass (2006-2007); species composition of degraded meadows and donor meadow were also collected (2004, 2006 and 2007). The soil seed bank (top 5 cm) at the two topsoil removal depths and seed content of hay were also sampled in 2004.


Referenced papers

Please cite as:

Dicks, L.V., Ashpole, J.E., Dänhardt, J., James, K., Jönsson, A., Randall, N., Showler, D.A., Smith, R.K., Turpie, S., Williams D.R. & Sutherland, W.J. (2017) Farmland Conservation Pages 245-284 in: W.J. Sutherland, L.V. Dicks, N. Ockendon & R.K. Smith (eds) What Works in Conservation 2017. Open Book Publishers, Cambridge, UK.