Employ areas of semi-natural habitat for rough grazing (includes salt marsh, lowland heath, bog, fen)
Overall effectiveness category Awaiting assessment
Number of studies: 9
Background information and definitions
Low intensity, rough grazing can be used to maintain open habitats in an early successional state. It can be enclosed (with limited stocking levels) or unenclosed. This action involves grazing semi-natural habitats that are not an integral part of modern farming practice, such as areas of peat bog, or other undrained wetland, lowland heath, saltmarsh and sand dunes. It does not include management of species-rich grasslands, wet meadows or wood pasture. For studies on the use of grazing to maintain these habitats, see “Agriculture and aquaculture – Maintain species-rich, semi-natural grassland”, “Agriculture and aquaculture – Maintain or restore traditional water meadows and bogs”, “Agriculture and aquaculture – Maintain or restore native wood pasture and parkland”.
For other management options for open, semi-natural habitats, see “Change mowing regime on grassland”, “Manage heathland by cutting” and “Natural system modifications – Use prescribed fire to maintain or restore disturbance in grasslands or other open habitats”.
Supporting evidence from individual studies
A controlled, before-and-after study in 1971–1973 in a fen in Cambridgeshire, UK (Duffey 1977) reported that after grazing by cattle, large copper butterflies Lycaena dispar batava laid more eggs/plant than either before grazing or than in ungrazed fens. Results were not tested for statistical significance. In the first two years after grazing commenced, the number of eggs (2.1–3.1 eggs/plant) was higher than in either the year before grazing on the same fen (0.1 eggs/plant) or in two ungrazed fens (0.1–1.7 eggs/plant). From late May–early August 1972–1973, one 4.2-ha fen was grazed by six bulls for 9 weeks/year, while two adjacent fens (2.3–3.3 ha) were not grazed. In summer 1972, a total of 137 male and 65 female adult butterflies were released from two cages, one in each ungrazed fen, to supplement the local population. In summer 1973, another 93 males and 70 females were released in one of the ungrazed fens. In August 1972, the vegetation height in the grazed fen (0.6–1.0 m) was lower than in the ungrazed fens (1.2–2.0 m). In the first week of August 1971–1973, all great water dock Rumex hydrolapathum plants in each site were examined, and the number of eggs counted.Study and other actions tested
A replicated, site comparison study in 1992–1996 in 22 calcareous coastal dunes in the Netherlands (WallisDeVries & Raemakers 2001) found that grazed sites had a higher abundance of some butterfly species than unmanaged sites or areas managed by cutting, but management type did not affect species richness. In 1996, the abundance of two out of 13 species was higher at grazed sites (small copper Lycaena phlaeas: 14 individuals/site; Queen of Spain fritillary Issoria lathonia: 62 individuals/site) than at unmanaged (small copper: 4; Queen of Spain fritillary: 23 individuals/site) or cut sites (small copper: 1; Queen of Spain fritillary: 4 individuals/site). Two other species were more abundant at grazed (small tortoiseshell Aglais urticae: 4; painted lady Vanessa cardui: 22 individuals/site) or unmanaged sites (small tortoiseshell: 5; painted lady: 22 individuals/site) than at sites managed by cutting (small tortoiseshell: 1; painted lady: 2 individuals/site). Brown argus Aricia agestis occurred more frequently in grazed (34% of sites) than in unmanaged (19%) or cut (11%) sites. The remaining eight species had similar abundances in grazed, unmanaged and cut sites (data not presented). Species richness was also similar between grazed (17 species/site), unmanaged (17 species/site) and cut (15 species/site) areas. Over four years, the total abundance of the 20 most common butterflies (out of 35 recorded) increased in grazed and unmanaged sites, but decreased in sites managed by cutting (data not presented). Eleven coastal dunes had been grazed year-round by cattle or ponies at low density (0.05–0.26 animals/ha/year) since 1983–1992, an additional four dunes were cut once/year in late July and a further seven were unmanaged. From April–October 1992–1996, butterflies were surveyed weekly along a 1-km transect at each site.Study and other actions tested
A replicated, site comparison study in 1998–1999 on 68 wet heathland sites in the Netherlands (WallisDeVries 2004) found that sites grazed year-round by horses and sheep had higher Alcon large blue Maculinea alcon occupancy than sites under other grazing regimes or ungrazed sites. Alcon large blue occupancy was higher in grazed plots (68%) than in ungrazed plots (41%), but was lowest in plots where grazing was combined with sod cutting (26%). Among grazed plots, occupancy was highest under year-round grazing by horses and sheep (77%), intermediate under year-round grazing by horses or sheep with summer grazing by cattle (56%), and lowest under summer grazing by cattle and sheep (29%). Sixty-eight wet heathland sites in the Netherlands where Alcon large blue was known to have occurred since 1990 were selected. Management information for the last five years was obtained by sending questionnaires to land managers. Grazing had been used at 44% of sites, with different livestock and regimes, but always equivalent to about 50 kg/ha/year (further details not provided). From mid-July–early September 1998–1999, Alcon large blue eggs were counted in each of three 10 × 10 m plots/site to determine butterfly presence in the plot.Study and other actions tested
A replicated, site comparison study in 2005–2007 at an upland site in the UK (exact location not given) (Fraser et al. 2008) found that cattle-grazed semi-natural upland rough grassland had a higher abundance and species richness of butterflies than permanently or partially grazed improved pasture. In semi-natural rough grassland, the abundance (905–1,938 individuals) and species richness (15–17 species) of butterflies was higher than either permanently grazed (abundance: 42–156 individuals; richness: 7–11 species) or partially grazed (abundance: 15–67 individuals; richness: 5–10 species) improved pasture. Eight butterfly species were found exclusively on the semi-natural grassland. Six semi-natural plots dominated by purple moor grass Molinia caerulea were grazed from June–September. Ten plots of improved perennial rye grass Lolium perenne/white clover Trifolium repens were grazed throughout the growing season by livestock. Ten similar plots were grazed in spring and autumn, but had livestock excluded from May–September and one silage cut taken. Butterfly transect counts were conducted weekly from mid-April to mid-September 2005–2007.Study and other actions tested
A controlled study in 1991–2009 in a saltmarsh in Schleswig-Holstein, Germany (Rickert et al. 2012, same experimental set-up as 6 and 9) found that lightly grazed and ungrazed saltmarsh supported more micro-moths than more intensively grazed saltmarsh. After 15–18 years of grazing, both the abundance and species richness of moths on a lightly grazed (abundance: light trap: 65.5, emergence trap: 7.1 individuals/trap; richness: light trap: 7.7, emergence trap: 1.3 species/trap) and an ungrazed marsh (abundance: light trap: 88.6, emergence trap: 6.0 individuals/trap; richness: light trap: 6.2, emergence trap: 1.8 species/trap) were higher than on a moderately grazed (abundance: light trap: 25.4, emergence trap: 2.3 individuals/trap; richness: light trap: 3.8, emergence trap: 0.6 species/trap) or heavily grazed marsh (abundance: light trap: 9.0, emergence trap: 0.2 individuals/trap; richness: light trap: 1.5, emergence trap: 0.1 species/trap). In 1991, four paddocks were established on a 1,050-ha saltmarsh and assigned to four grazing treatments: light (1–2 sheep/ha), moderate (3–4 sheep/ha) or heavy grazing (10 sheep/ha), and ungrazed (0 sheep/ha). From June–September 2006–2009, micro-moths were sampled using one 12 V actinic light trap/paddock on 6–9 nights/year (31 nights total). From April–October 2007–2009, micro-moths were sampled using a 1-m2 steel emergence trap in each of three 150 × 20 m plots/paddock (>250 m apart). Traps were emptied weekly and repositioned every three weeks, therefore sampling 10 m2/plot/year.Study and other actions tested
A replicated, paired, controlled study in 1988–2010 in three saltmarshes in Schleswig-Holstein, Germany (van Klink et al. 2013, same experimental set-up as 5 and 9) reported that lightly grazed saltmarsh had a higher abundance of two moth caterpillars, but a lower abundance of four caterpillars, than ungrazed saltmarsh, and heavily grazed marsh had a lower abundance of all species. Results were not tested for statistical significance, and data were not presented. In lightly grazed saltmarsh, two caterpillars (wormwood case-bearer Coleophora artemisiella and saltern-rush case-bearer Coleophora adjunctella) were more abundant than in heavily grazed or ungrazed marsh. Four species (maritime bell Eucosma lacteana, saltmarsh bell Eucosma tripoliana/pale saltern bell E. rubescana, common sea groundling Scrobipalpa nitentella, netted bagworm Whittleia retiella) were more abundant in ungrazed marsh than in grazed marshes, and one species (saltmarsh case-bearer Coleophora atriplicis) occurred in similar numbers in all marshes. None of the seven species recorded were more abundant in heavily grazed saltmarsh than in lightly grazed or ungrazed marsh. In 1988–1991, three heavily grazed (10 sheep/ha) saltmarshes were each divided into three paddocks (11–15 ha) and assigned to either heavy (10 sheep/ha), light (3–4 sheep/ha) or no (0 sheep/ha) grazing. The marshes were grazed from May–October, with sheep removed only during flooding events. In September 2010, moth caterpillars were sampled in sixteen 30-cm diameter points/paddock. In lightly grazed paddocks, eight points were in taller vegetation and eight in short vegetation. Caterpillars were collected by suction sampling (30 seconds), followed by removing and sieving all vegetation above 3 cm, and another 30-second suction sample.Study and other actions tested
A replicated, randomized, paired, controlled study in 2007–2008 in a meadow steppe grassland in Jilin Province, China (Zhu et al. 2015) found that moderately grazed plots had a lower abundance of butterflies and moths than ungrazed plots. After a year and a half of grazing, the abundance of butterflies and moths on plots grazed by cattle (2–7 individuals/plot), goats (3–7 individuals/plot) or sheep (3–6 individuals/plot) was lower than on ungrazed plots (6–22 individuals/plot). Nine 0.3-ha blocks were each divided into four fenced, 0.05-ha plots, 18–20 m apart, to which four grazing treatments were randomly assigned. From July 2007 and 2008, plots were either grazed by two cattle, eight goats, or eight sheep, or left ungrazed. Grazing was conducted for two hours each morning and evening, until 60% of forage was removed (10–15 days/month, number of months not given). From July–October 2008, insects were surveyed four times by walking two 25-m-long transects/plot, twice/day, and taking 15 sweeps/transect through the vegetation with a 40-cm diameter net. All adult insects were identified to species.Study and other actions tested
A replicated, paired, site comparison study in 2010 in eight shrubsteppe sites in British Columbia, Canada (Elwell et al. 2016) found that grazed shrubland did not support a higher abundance, species richness or diversity of pollinators (including butterflies) than ungrazed shrubland. On grazed shrubland, the total abundance (469–1,188 individuals), species richness (82–124 species) and diversity (data presented as model results) of pollinators, including butterflies, were all similar to ungrazed shrubland (abundance: 576–925 individuals; richness: 86–113 species). Four pairs of sites (20–1,850 ha), similar in topography and vegetation, were selected. Within each pair, one site was grazed with cattle for 4–6 weeks between April and June annually or biannually, at 14–160 cows/ha/month, while the other site had been ungrazed for 6–40 years. From April–July 2010, pollinating insects were sampled eight times (bi-weekly) by setting 30 yellow, white and blue 12-oz pan-traps for 8.5 hours at 3-m intervals diagonally across a 1-ha plot at each site. Paired sites were always sampled on the same day.Study and other actions tested
A controlled study in 1991–2009 in a saltmarsh in Schleswig-Holstein, Germany (Rickert et al. 2018, same experimental set-up as 5 and 6) found that a lightly grazed and an ungrazed saltmarsh had a different species community of micro-moths compared to a heavily grazed marsh. After 16–18 years of grazing, the moth community in a lightly grazed and an ungrazed saltmarsh was different from that in a heavily grazed saltmarsh (data presented as model results). See paper for individual species results. Until 1990, a 1,050-ha saltmarsh was grazed at 10 sheep/ha. From 1991, four paddocks (100–256 ha) were lightly (1–2 sheep/ha), moderately (3–4 sheep/ha) or heavily grazed (10 sheep/ha), or left ungrazed (0 sheep/ha). The marsh was grazed from May–October, with sheep removed only during flooding events. From April–October 2007–2009, micro-moths were sampled using a 1-m2 steel emergence trap in each of three 150 × 20 m plots/paddock (>250 m apart). Traps were emptied weekly and repositioned every three weeks, therefore sampling 10 m2/plot/year.Study and other actions tested