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: Reduce intensity of grazing by domestic livestock Terrestrial Mammal Conservation

Key messages

Read our guidance on Key messages before continuing

  • Thirteen studies evaluated the effects on mammals of reducing the intensity of grazing by domestic livestock. Six studies were in the USA, six were in Europe and one was in China.

COMMUNITY RESPONSE (3 STUDIES)

  • Richness/diversity (3 studies): Two of three site comparison or controlled studies, in the USA and Norway, found that reduced livestock grazing intensity was associated with increased species richness of small mammals whilst one study did not find an increase in species richness.

POPULATION RESPONSE (13 STUDIES)

  • Abundance (13 studies): Six of nine site comparison or controlled studies (including seven replicated studies), in the USA, Denmark, the UK, China, Netherlands and Norway, found that reductions in livestock grazing intensity were associated with increases in abundances (or proxies of abundances) of small mammals, whilst two studies showed no significant impact of reducing grazing intensity and one study showed mixed results for different species. Two replicated studies (including one controlled and one site comparison study), in the UK and in a range of European countries, found that reducing grazing intensity did not increase numbers of Irish hares or European hares. A controlled, before-and-after study, in the USA found that exclusion of cattle grazing was associated with higher numbers of elk and mule deer. A replicated, site comparison study in the USA found that an absence of cattle grazing was associated with higher numbers of North American beavers.

BEHAVIOUR (0 STUDIES)

Supporting evidence from individual studies

1 

A controlled, before-and-after study in 1981–1982 in a forest and meadow mosaic in Arizona, USA (Wallace & Krausman 1987) found that an absence of cattle grazing was associated with higher numbers of elk Cervus canadensis and mule deer Odocoileus hemionus. There were 0.13 elk/km counted on transects in absence of cattle grazing and 0.01/km after grazing commenced whereas, concurrently, on a continually ungrazed pasture, 0.21 and 0.50 elk/km respectively were counted. The number of mule deer counted on transects fell from 0.07/km in absence of grazing to 0.00/km after grazing commenced whereas 0.02 mule deer/km were counted on a continually ungrazed pasture during both time periods. The 135 km2-study area was divided into two pastures. One was ungrazed in both years. The other was ungrazed in 1981 and stocked with cattle, at a rate of one animal unit (equivalent to a cow and suckling calf)/3 ha in May–July 1982. Elk and mule deer were counted in July and August, along a 48-km driving transect, 20 times in 1981 and 14 times in 1982.

2 

A site comparison study in 1981–1983 on a grassland ranch in Arizona, USA (Bock et al. 1984) found that reducing grazing intensity by excluding livestock increased rodent abundance. More rodents were caught in an ungrazed area (428 individuals) than in a grazed area (328 individuals). This was the case for hispid pocket mouse Perognathus hispidus (38 vs 16 individuals), western harvest mouse Reithrodonromys megalotis (26 vs 4), white-footed mouse Peromyscus leucopus (45 vs 24), southern grasshopper mouse Onychomys torridus (42 vs 8) and hispid cotton rat Sigmodon hispidus (118 vs 49). Merriam’s kangaroo rat Dipodomys merriami was less abundant in the ungrazed than the grazed area (5 vs 92 individuals). Silky pocket mouse Perognathus flavus abundance did not differ significantly between ungrazed and grazed areas (8 vs 5 individuals) and nor did deer mouse Peromyscus manicularus abundance (146 vs 130). Livestock were fenced out of part of a 300-ha study area from 1968 onwards. The grazed part was stocked with approximately one cow/10 ha. Rodents were live-trapped, from two hours before sunset to two hours after sunrise, on 71 occasions, from July 1981 to January 1983.

3 

A replicated, site comparison study in 1989–1991 of shrub grassland in a national park in Utah, USA (Rosenstock 1996) found that reducing grazing intensity by excluding cattle from small enclosures did not increase small mammal abundance or species richness. Small mammal abundance in ungrazed enclosures (1.9 individuals/100 trap-nights) did not significantly differ from that in grazed areas (2.3 individuals/100 trap-nights). Small mammal species richness in enclosures (1.5 species/trap grid) did not significantly differ from that in grazed areas (1.6 species/trap grid). Cattle were excluded from four enclosures, three for six years prior to the study and one for 38 years. Enclosures measured 0.1–0.8 ha. Grazing outside enclosures was by 1,500 Animal Units (equivalent to a cow and suckling calf) across 35,499 ha in October–May. Small mammals were sampled in grids of Sherman live traps, one grid inside each enclosure. An identical grid was sampled simultaneously >500 m away from each enclosure. Grids were trapped for four consecutive days, between 1 May and 31 June. Three enclosures were sampled annually in 1989–1991, and one in 1990–1991.

4 

A replicated, site comparison study in 1990 of shrub grassland at eight sites in two national parks in Utah, USA (Rosenstock 1996) found that reducing grazing intensity by excluding cattle from areas of grassland increased small mammal abundance and species richness. Small mammal abundance in ungrazed sites (1.8 individuals/100 trap-nights) was higher than in grazed sites (1.0 individuals/100 trap-nights). Small mammal species richness in ungrazed sites (1.5 species/site) was higher than in grazed sites (1.0 species/site). Eight sites were sampled; four ungrazed for ≥30 years and four in a region grazed by 1,500 Animal Units (equivalent to a cow and suckling calf) across 35,499 ha in October–May. All sites were on large (≥ 100 ha) areas of shrub-grassland and were selected to match geological and soil characteristics. Each site was sampled using a grid of Sherman live traps, for four consecutive days, between 1 May and 31 June 1990.

5 

A replicated, controlled study in 1998–2000 of pasture at a site in Denmark (Schmidt et al. 2005) found that in plots with reduced livestock grazing intensity, small mammal biomass was higher. Small mammal biomass peaks across the study in each of two plots/treatment were higher in ungrazed plots (287–959 g), intermediate in low-intensity sheep plots (251–801 g) and lowest in high-intensity cattle plots (64–195 g). The estimated population of field voles Microtus agrestis (the most abundant species recorded) was higher each year in ungrazed plots (29–94/plot) than in high-intensity cattle plots (3–27/plot), but was higher still in low-intensity sheep plots in two of three years (32–63/plot). In 1997, two meadows were divided into 70 × 300-m pens. One plot on each meadow was assigned to high-intensity cattle grazing (4.8 steers/ha), one to low intensity sheep grazing (4.5 ewes plus lambs/ha) and one was ungrazed. Grazing occurred from mid-May to mid-October, though was prevented on half of each pen until after hay cutting (late-June to early-July). The delayed grazing part was reversed the following year. Small mammals were live-trapped over three days and nights, every four weeks, over 31 trapping sessions, from June 1998 to October 2000.

6 

A replicated, randomized, paired sites, controlled, before-and-after study in 2002–2004 on upland grassland in Scotland, UK (Evans et al. 2006) found that reducing sheep grazing intensity increased the abundance of field voles Microtus agrestis. In the first year of grazing treatments, the percentage of quadrats with vole signs was higher in ungrazed plots (20%), intermediate in lightly grazed plots (12%) and lowest in heavily grazed plots (4%). The same pattern held in the second year of treatments (ungrazed: 24%; lightly grazed: 11%; heavily grazed: 7%). Before grazing treatments were implemented, there was no significant difference in the frequency of vole signs between plots. Plots were all grazed similarly (stocking rate not stated) up to 2002. From spring 2003, there were six replicates (3.3 ha each) of no livestock grazing, light grazing (three ewes/plot) and heavy grazing (nine ewes/plot). Five 25 × 25-cm quadrats at each of five points/plot were searched for vole signs in April and October 2002–2004.

7 

A replicated, site comparison study in 2001 and 2002 on two winter pasture areas in Sichuan, China (Raoul et al. 2006) found that reduced livestock grazing intensity was associated with higher numbers of the tundra/lacustrine vole Microtus oeconomus/limnophilus complex but with lower numbers of Kam dwarf hamster Cricetulus kamensis. The numbers of tundra/lacustrine voles in low grazing intensity areas (7 individuals/100 trap nights) was higher than in medium (1/100 trap nights) or high grazing intensity areas (0/100 trap nights). The numbers of Kam dwarf hamster in low (0 individuals/100 trap night) and medium grazing intensity areas (0/100 trap nights) was lower than that in high grazing intensity areas (6/100 trap nights). Surveys were conducted in grassland and shrub areas in valley, wetland and slope habitats in winter pasture at 4,250 m altitude. Sites were grazed, in varying intensities, by yaks, sheep, goats, and horses, each October to early May. Small mammals were surveyed using back-break traps over three nights and days in July 2001 and July 2002.

8 

A replicated, site comparison study in 2005 on 200 plots covering a range of agricultural habitats in Northern Ireland, UK (Reid et al. 2007) found that reducing grazing intensity as part of a wider suite of agri-environment measures did not increase numbers of Irish hares Lepus timidus hibernicus. The effects of reducing grazing intensity cannot be separated from those of other agri-environment measures, which included retaining and enhancing field boundary features and managing nutrient systems. Hare abundance in agri-environment plots (0.45 hares/km transect) did not significantly differ from that in non-agri-environment plots (0.41 hares/km transect). One hundred and fifty 1-km2 plots, on land that was enrolled into an agri-environment scheme 10–17 years previously, were selected along with 50 non-enrolled 1-km2 plots, chosen to match enrolled plots for landscape characteristics. Hares were surveyed at night, in mid-winter, by spotlighting from a vehicle.

9 

A replicated, randomized, controlled study in 2002–2004 on grassland in France, Germany, Italy and the UK (Wallis De Vries et al. 2007) found that areas with low livestock grazing intensities did not have more European hares Lepus europaeus than did areas with moderate livestock grazing intensities. Too few hares were recorded to enable statistical analyses. At the UK site, though, where most hares were recorded, numbers were similar between low intensity (14 hares) and moderate intensity (12 hares) grazing areas. Sites were grazed by the cattle Charolais × Fresian in the UK, Simmental in Germany and Charolais in France and by Finnish Romanov sheep in Italy. Grazing rates differed, but low grazing intensity was 0.3–0.4 fewer animals/ha than moderate grazing intensity. There were three each of low and moderate intensity grazing paddocks (paddock size 0.4–3.6 ha) at one site in each of the four countries. Hares were counted every two weeks in early morning, from May to October, 2002–2004, during seven minutes of observation and whilst walking a transect in each paddock.

10 

A controlled study in 2008 of a grassland and woodland site in Nevada, USA (Rickart et al. 2013) found that reducing grazing intensity by long-term exclusion of domestic livestock resulted in a higher species richness and abundance of small mammals. More small mammal species were recorded on ungrazed land (six) than on grazed land (four). Small mammal abundance on ungrazed land (0.08 animals/trap night) was higher than on grazed land (0.05 animals/trap night). Three species were caught in sufficient quantities for individual analyses. The Great Basin pocket mouse Perognathus parvus was more abundant on ungrazed than grazed land (0.05 vs 0.02 individuals/trap night) as was western jumping mouse Zapus princeps (0.02 vs 0.00 individuals/trap night). Deer mice Peromyscus maniculatus showed no preference (0.01 vs 0.01 individuals/trap night). Sampling occurred in a 10-ha enclosure, characterised by mixed shrubs and trees, from which domestic livestock were excluded at least 50 years previously and in a similar sized, adjacent cattle-grazed grassland. Small mammals were sampled using lines of snap-traps, over three or four nights, in July 2008.

11 

A replicated, site comparison study in 2013 in a forested area in New Mexico, USA (Small et al. 2016) found that an absence of cattle grazing was associated with higher numbers of North American beavers Castor canadensis. The relative frequency of beaver dams was higher in the absence of cattle grazing than where cattle grazing was present (data presented as odds ratios). Data were collected along 57 sections of river, each 200 m long, of which 29 had beaver dams and 28 did not have beaver dams, though physical conditions were suitable for their construction. Field data were collected between 15 May and 15 August 2013. Livestock grazing was assessed by collating information on grazing consents and by surveying ungulate faeces.

12 

A replicated, randomized, paired sites, controlled study in 2010–2013 on a coastal salt marsh in the Netherlands (van Klink et al. 2016) found that plots grazed at lower intensity contained more signs of vole Microtus spp. presence than did plots grazed at higher intensity. After four years, a greater proportion of surveyed quadrats contained signs of vole presence in plots grazed at lower intensity than in plots grazed at high intensity (data not reported). Twelve plots were established (in three sets of four) on a historically grazed salt marsh. From 2010, six plots (two random plots/set) were grazed at each intensity: low (0.5 animals/ha) or high (1.0 animal/ha). Grazing occurred in summer (June–October) only. Half of the plots were grazed by cows and half by horses. In October 2013, sixty quadrats (2 m2) were surveyed in the higher elevations of each plot for signs of vole presence (runways, fresh plant fragments or faecal pellets). Some flooded quadrats were excluded from the analysis.

13 

A randomized, replicated, controlled study in 2002–2005 at two heathland sites in Norway (Spirito et al. 2017) found that excluding livestock with fences did not significantly change abundances of field voles Microtus agrestis. The number of animals trapped in plots that were fenced to exclude livestock did not differ significantly (6 animals/plot) from that in plots that were not fenced to exclude livestock (4 animals/plot). In 2002, at two sites, four 50 × 50-m plots were fenced to exclude livestock and four plots were not fenced. Sheep density prior to fencing was 32–48 sheep/ha. In June and August 2003–2005, thirty-six live traps baited with sunflower seeds and peanuts and with wool for bedding were placed in each plot and checked twice daily for five days. Captured animals were individually marked and released.

Referenced papers

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.