Providing evidence to improve practice

Action: Other biodiversity: Exclude grazers Mediterranean Farmland

Key messages

Amphibians (1 study): One replicated, randomized, controlled study in wet grasslands in the USA found no difference in the abundance of Yosemite toads between areas with cattle excluded and grazed areas.

Birds (2 studies): One replicated site comparison in desert in the USA found more bird species, and more species that were nesting, in areas with sheep excluded, compared to grazed areas. Two replicated site comparisons in desert and wetlands found higher abundances of some or all species of birds in areas with cattle or sheep excluded, compared to grazed areas. The wetland study also found lower abundances, in some comparisons.

Fish (2 studies): One replicated site comparison in grasslands in the USA found higher biomass and abundance of golden trout in areas with cattle excluded, compared to grazed areas. Another one found fewer trout nests in part of a stream with a livestock exclosure, compared to part without a livestock exclosure.

Invertebrates (5 studies): Two replicated studies (one randomized and controlled) in wetlands and grasslands in the USA found more species or families of invertebrates in areas with cattle excluded, compared to grazed areas, for some or all groups. One replicated, randomized, controlled study in grasslands in the USA found fewer aquatic invertebrate species in areas with cattle excluded, compared to grazed areas, in some comparisons. Two replicated studies (one randomized and controlled) in grasslands in the USA found no difference in invertebrate abundance between ungrazed and cattle-grazed plots. One replicated, before-and-after site comparison in grasslands in the USA found that populations of a threatened, endemic butterfly declined in sites with cattle excluded, but also declined in cattle-grazed sites.

Mammals (4 studies): Two replicated site comparisons in deserts and grasslands in Spain and the USA found more mammal species in areas with cattle or sheep excluded, compared to grazed areas. One of these studies also found higher mammal diversity, and both studies found higher mammal abundance, in areas with grazers excluded, compared to grazed areas, in some or all comparisons. One replicated site comparison in desert in the USA found lower abundances of black-tailed hares in ungrazed sites, compared to grazed sites, and one replicated, randomized, controlled study in wooded grassland in the USA found no difference in ground squirrel abundance between ungrazed plots and cattle-grazed plots.

Plants (41 studies)

  • Abundance (38 studies): Thirty-two studies (13 replicated, randomized, and controlled) in grasslands, shrublands, wetlands, deserts, and mixed habitats in the USA, Israel, Chile, Spain, and Australia found higher biomass, cover, or abundance of some or all plant groups (or lower cover of non-native species), in areas with cattle, sheep, goats, or alpacas excluded, compared to grazed areas, in some or all comparisons. Fourteen studies (four replicated, randomized and controlled) from the USA, Israel, Spain, and Australia found lower biomass, cover, or abundance of some or all plant groups (or higher cover of non-native species), in areas with grazers excluded, compared to grazed areas, in some comparisons. Five replicated, controlled studies (four randomized) in grasslands in the USA found no difference in the cover of plants (and/or non-native plants) between ungrazed and grazed areas.
  • Diversity (19 studies): Five studies (three replicated) in forests, shrublands, and grasslands in Israel, Spain, and the USA found more species, or fewer non-native species, in areas with cattle or sheep excluded, compared to grazed areas, in some or all comparisons. Nine studies in grasslands and shrublands in Australia, Israel, Spain, and the USA found fewer species or native species, larger decreases in the number of species, or smaller increases in the number of species, in areas with cattle, sheep, or alpacas excluded, compared to grazed areas, in some or all comparisons. Six studies in grasslands, wetlands, and deserts in the USA found no differences in the number of species between areas grazed by cattle, sheep, or alpacas, and ungrazed areas. Four studies in shrublands, grasslands, and wetlands in the USA and Israel found higher plant diversity, or different community composition, in plots with cattle excluded, compared to grazed plots, in some comparisons. Three studies in wetlands and grasslands in the USA found lower plant diversity in plots with cattle excluded, compared to grazed plots, in some comparisons. Three studies in deserts and shrublands in the USA and Israel found no difference in plant diversity between plots with cattle or sheep excluded and grazed plots.
  • Survival (2 studies): One replicated, randomized, controlled study along creeks in the USA found that similar percentages of planted willows survived in pastures with or without cattle excluded. One replicated, randomized, controlled study in grasslands in the USA found higher plant survival in plots with cattle excluded, compared to grazed plots, in some comparisons.

Reptiles (1 study): One replicated site comparison in desert in the USA found lower abundances of reptiles, and of some reptile species, in areas with sheep excluded, compared to grazed areas, in some comparisons.

Implementation options (1 study): One site comparison in the USA found that more plant species were found in historically cultivated sites that were ungrazed, compared to grazed, but similar numbers of plant species were found in historically uncultivated sites that were ungrazed or grazed.

Supporting evidence from individual studies

1 

A replicated, randomized, controlled study in 1982–1984 in alpine meadows in central California, USA, found higher biomass of some plant groups in plots from which cattle were excluded, compared to grazed plots. Plants: The peak biomass of non-grass plants was higher in plots with cattle excluded, compared to grazed plots, in one of two meadows, one year after fertilizer was added (252 vs 99–138 g/m2). However, there were no differences in the other meadow, in the other two years, or when plots were not fertilized (30–249 g/m2). There was no difference in the biomass of grass between ungrazed and grazed plots for either meadow, in any year (170–480 g/m2). There were no differences between ungrazed and grazed plots in the cover of sedges and rushes (7–71% cover) or non-grass plants (20–71% cover). Methods: Eighteen plots were established in two grazed meadows in 1982, with cattle excluded from half. The vegetation in plots was sampled at 30 points in July and August 1982–1984. Half of the plots were also fertilized in 1982.

 

2 

A controlled study in 1983–1985 in the central Sierra Nevadas, California, USA, found more herbaceous vegetation and more vegetation cover in plots with cattle excluded, compared to grazed plots. Plants: At the end of the growing season, more herbaceous vegetation was found in ungrazed, compared to grazed plots (1,580 vs 150–760 kg/ha). By the end of the grazing season, cover of vegetation less than 50 cm tall was higher in ungrazed, compared to grazed plots (49–85% vs 17–68% cover), although these differences were present before cattle were introduced, in one of three vegetation types. Cover of vegetation less than 1 m tall was higher in ungrazed plots, in one of three vegetation types (75% vs 52–54%). Taller vegetation did not differ between ungrazed and grazed plots (10–72%). Methods: Three plots (22–29 ha) were fenced in 1983, and were grazed in 1984–1985 at one of three levels (no cattle, moderate grazing, or heavy grazing) for 48–74 days. Densities were 0.65–0.76 and 1–1.7 animal unit months/ha, respectively, and plots were grazed for up to 100 days each year. Each plot received a different treatment in each year. Vegetation cover was monitored in quaking aspen Populus tremuloides (all years), willow Salix sp., and corn lily Veratrum californicum (1984–1985) throughout each grazing season.

 

3 

A replicated, randomized, controlled study in 1987–1988 in pastures along three creeks in the northern Sierra Nevada, California, USA, found that Geyer willow Salix geyeriana plantings had similar survival rates in pastures without or with cattle, but fewer were grazed in pastures without cattle. Plants: Similar percentages survived in pastures without or with cattle (33% vs 18–26%), but fewer were grazed in pastures without cattle (0.2 vs 0.7–1). Methods: The cuttings (over two-years old, 10.5 mm diameter, 42.3 cm length) were planted (30 cm depth) along three creeks (Cow, Freeman, and Big Grizzly), on thirty transects (3 m apart) across each creek and extending 10 m from the top of each bank, in May 1987 (300 plantings/creek). Along each creek, three pastures were fenced: one pasture was ungrazed by cattle, one was grazed early (1987: 22 June–7 August; 1988: 21 June–1 July) and one was grazed late (1997: 4 August–23 September; 1988: 18 August–31 August). The cuttings were measured in September 1987 and 1988.

 

4 

A replicated, controlled study in 1982–1985 in mixed savanna, shrubland, and grassland in northern California, USA, found more plant species and higher plant cover in ungrazed plots, compared to sheep-grazed plots. Plants: More plant species were found in ungrazed plots, compared to grazed plots, in autumn-winter grazed pastures in woodland (8 vs 7 species/50 sample points), but there were no differences with spring grazing or in grassland (data not reported). Plant cover was higher in ungrazed plots, compared to grazed plots, within spring grazed pastures in both woodland and grassland (75–88% vs 67–76% cover), and in autumn-winter grazed pastures in woodland, but not grassland (65% vs 58%). Seven of 15 species of plant had different amounts of cover in ungrazed, compared to grazed plots, in one of four habitat-grazing combinations each time. Methods: Two pastures were established in mixed blue oak Quercus douglasii woodland and grassland areas in 1982 and were grazed by sheep from May until October each year, with autumn or spring grazing. Plants were monitored throughout the year in 20 plots within both woodland and grassland in each pasture, with 75 x 75 cm cages to exclude grazers.

 

5 

A replicated, randomized, controlled study in 1989–1991 in grasslands in north-central California, USA, found few differences in the survival of, or damage to, blue oak Quercus douglasii seedlings planted in pastures from which cattle were excluded, compared to grazed pastures. Plants: There were no differences in 15-month survival between seedlings planted in ungrazed plots, compared to those grazed by cattle in spring or summer (9–24% survival). Seedlings had lower survival in ungrazed plots, compared to winter grazed plots (15% vs 46%). The proportion of seedlings damaged by browsing or trampling did not differ between ungrazed and grazed plots (0–85% damaged). Methods: In December 1989, oak seedlings were planted in three pastures, each containing ten plots: one with cattle excluded and nine grazed for a week each at different intensities and at different times. Ungrazed plots were accessible to wild herbivores. Each plot received 24 seedlings (720 in total), of which half had the area around them treated with glyphosate herbicide to reduce competition from grass.

 

6 

A replicated site comparison in 1990–1992 in a desert in south-central California, USA, found more species, and higher densities, of some small mammal species, in plots from which sheep were excluded, compared to grazed plots. More plant biomass was found in ungrazed plots, compared to grazed plots. Plants: Similar plant diversity was found in ungrazed and grazed plots (3.5–4.3 vs 1.5–3.7 species; diversity reported as diversity indices). More plant biomass was found in ungrazed plots (12–199 vs 5–57 kg/ha). Plant biomass was higher in ungrazed areas, in 21 of 23 species, in at least one year, and the two species with lower biomass were non-native. The cover of two of 13 species of perennial shrub was higher in ungrazed areas (Ambrosia dumosa: 2.8% vs 1.5% cover; Lycium andersonii: 0.9% vs 0.1%). The density of one of 13 species of perennial shrubs was lower in ungrazed areas (Acamptopappus sphaerocephalus: 85 vs 333 plants/ha). The density of seeds was higher in ungrazed plots, in one of three years (1992: 193 vs 56 kg/ha; other years: 50–121). The biomass of non-grasses was higher than that of non-native annual grasses in all years in the ungrazed area, but only in one year in the grazed area. Mammals: More species of small nocturnal rodents were found in ungrazed areas (3.7 vs 2.5 species/sample), and diversity was higher in ungrazed areas in all three years (data reported as diversity indices). The densities of three of five small, nocturnal rodents were higher in ungrazed plots (long-tailed pocket mouse: 26 vs 6 animals/ha; Merriam’s kangaroo rat: 31 vs 13; southern grasshopper mouse: 3 vs 0). Methods: Two pairs of 65 ha plots were established in 1990 (one plot inside an area fenced since 1978–1979 and one outside). Vegetation was sampled in ten 1 m2 plots and at ten points on each of ten 100 m transects in April and June 1990–1992. Seeds were sampled in the top 2 cm of the soil in ten 6 cm diameter circles. Mammals were caught in 64 Sherman traps in each plot (in five periods of 4–6 nights).

 

7 

A replicated site comparison in 1993–1994 in alpine meadows in central California, USA, found more golden trout Oncorhynchus mykiss aguabonita in streams, and more willows on stream banks, in areas of meadows from which cattle were excluded, compared to grazed areas. Ungrazed plots also contained larger willows and more young willows. Fish: More golden trout were found in streams in ungrazed plots, compared those in three of four grazed plots (1.4–2.7 vs 1.3–2.2 fish/m2). Golden trout biomass was higher in streams in ungrazed plots, compared those in three of four grazed plots (20–21 vs 16–18 g/m2). Plants: Ungrazed plots contained more willows than grazed plots (124–246 vs 11–70 trees/125 m of bank), and contained larger willows than three of four grazed plots (140–220 vs 20–70 cm height for largest tree). Young willows less than 20 cm were more abundant in ungrazed plots (16–134 vs 1–8 trees/125 m of bank). Canopy shading over streams was higher in ungrazed plots, for three of four plots (32–35% vs 2–24% shading). Methods: Fences were erected in 1983 and 1991 in two meadows to exclude cattle from a total of three areas. Fish and vegetation were monitored in 125 m sections either upstream or downstream of the exclosures and inside them (a total of seven sites, three ungrazed) in August 1993–1994. Fish were surveyed by electrofishing and vegetation using transects every 5 m along the stream. Areas outside the exclosures were grazed by cattle in July and September.

 

8 

A site comparison in 1991 in annual grassland on the Central Coast, California, USA, found more plant species and different plant communities in ungrazed grassland, compared to grazed grassland, after over 50 years of grazer exclusion. Plants: More plant species were found in ungrazed sites, compared to grazed sites (33 vs 27 species), and there were differences in the plant communities (reported as differences in ordination space). The invasive, non-native medusahead grass Elymus (Taeniatherum) caput-medusae was found only in grazed sites, but the native grass Elymus glaucus was found only in ungrazed sites (see publication for details of other species). Implementation options: When split into cultivated or uncultivated sites, more plant species were found in historically cultivated sites that were ungrazed, compared to grazed (32 vs 24 species), but similar numbers of plant species were found in uncultivated sites that were ungrazed or grazed (33 vs 30). Methods: European domestic cattle were introduced to Monterey County in 1770. In 1937, grazers were excluded from one landscape (the Hastings Natural History Reservation), but not from a nearby landscape. In 1991, 43 sites in the ungrazed landscape and 37 sites in the grazed landscape were sampled (methods not clearly reported, but plant cover was measured in 20 x 50 m plots in April–May in a different part of this study).

 

9 

A replicated site comparison in 1994 in meadow streams in Inyo National Forest, California, USA, found fewer California golden trout Oncorhynchus mykiss aguabonita nests in part of a stream with a livestock exclosure, and fewer nests in narrower, deeper streams. Fish: Fewer trout nests (redds) were found in part of stream with a livestock exclosure, compared to part without a livestock exclosure (0 vs 52 nests; 0 vs 0.42 nests/m2), in the Upper Ramshaw Creek. In Mulkey Creek, fewer trout nests were found in narrow streams, compared to wide streams (0 vs 0.27 nests/m2), and grazer exclusions were associated with narrower streams in this area, in previous studies. Methods: Trout nest density was measured in narrow (45–125 m) and wide (53–130 m) stretches of streams (five replicates each) in Mulkey Creek. In the Upper Ramshaw Creek, nest density was measured inside (upstream) and outside (downstream) of a livestock exclosure. Nests were surveyed six times between 1 May and 5 June.

 

10 

A replicated site comparison in 1994–1996 in a desert site in south-central California, USA, found more birds and bird species in plots with grazers excluded, compared to sheep-grazed plots. Fewer black-tailed hares Lepus californicus, but more lizards were found in fenced plots, compared to unfenced plots. Perennial plant cover was higher in fenced plots, compared to unfenced plots. Birds: More bird species were observed nesting in fenced plots, compared to unfenced plots (3 vs 1 species). More birds and bird species were found in fenced plots (0.9–3.1 vs 0.7–2.6 species/survey; 1–11 vs 1–9 birds/survey), and six of 22 species were more abundant in fenced plots. Mammals: Fewer black-tailed hares were found in fenced plots (0–1.5 vs 1–4 hares/survey; 11 vs 22–31 droppings/m2). Plants: There were no differences in species diversity of perennial plants in fenced or unfenced plots (data reported as Shannon-Weiner indices). Perennial plant cover was higher in fenced plots (13–14% vs 6–7% cover). There were no differences in diversity of height, cover, or volume of perennial plants between fenced and unfenced sites (data reported as Shannon-Weiner indices). Reptiles: Fewer lizards were found in ungrazed plots, compared to grazed plots (1–4 vs 2–10 lizards/survey), and two of six species were less abundant in ungrazed plots, in some comparisons. Methods: Two 2.25 ha plots that were fenced in 1980 were compared to two plots that were grazed by sheep until 1994. Sites were matched for environmental variables. Birds were counted using 16 point counts in each plot, four times during breeding seasons (1994–1995) and twice during winter (December 1994, January 1996). Lizards were surveyed using 1.25 km transects three times in summer (1994–1995). Hares numbers were estimated with four 1.25 km transects and in sixty 40 x 50 cm sampling units in each plot. Plants were surveyed at 16 points in each plot in June 1995. Unfenced plots were also driven over by off-highway vehicles.

 

11 

A replicated, randomized, controlled study in 1989–1999 in coastal grasslands in central California, USA, found that the occurrence of native grasses declined by more in plots with grazers excluded, compared to cattle-grazed plots, but different species responded in different ways. Plants: The occurrence of native grasses declined by more in plots with grazers excluded than in grazed plots (20% vs 3% decline). Percentage cover of native grasses did not differ between plots with grazers excluded and grazed plots (11% decrease to 5% increase vs 3–8% increase). Cover of one of three native grass species (Danthonia californica) decreased by 12% on ungrazed plots, but increased by 10% on grazed plots. Occurrence of D. californica was affected by the plot’s location, decreasing by more in ungrazed plots than in grazed plots on lower slopes (58% vs 8% decline), but not on middle or upper slopes (1–16% increase). Occurrence of Nassella lepida decreased by more on ungrazed, compared to grazed, upper slopes (10% vs 3% decrease), but increased more on ungrazed, compared to grazed, lower slopes (5% increase vs 13% decrease). Nassella pulchra occurrence decreased by more on ungrazed, compared to grazed, upper slopes (20 vs 3% decrease). Methods: Three 0.25 ha areas were established in 1989 in the upper, middle, and lower slopes of the site, with grazers excluded from one plot in each area. The remaining area was grazed by cattle and sheep. The cover and occurrence of native grasses was assessed in spring 1989 and 1991.

 

12 

A replicated, before-and-after site comparison in 1991–1998 in serpentine grasslands in the San Francisco Bay Area, California, USA, found that populations of the threatened, endemic, Bay checkerspot butterfly Euphydryas editha bayensis declined in sites with cattle excluded, but also declined in cattle-grazed sites. One host plant of this butterfly had lower cover in sites with cattle excluded, compared to cattle-grazed sites, and two non-native grasses had higher cover. Invertebrates: Numbers of Bay checkerspot butterfly larvae first increased, but then decreased to local extinction, after cattle were excluded from two sites in Silver Creek (excluded from 1989 and 1992 onwards, respectively; maximum: 75,000 larvae in 1993; minimum: 0 in 1995–1997). However, decreases were also seen in 1994–1997 at nearby sites from which cattle were not excluded (Coyote Ridge, maximum: 38,000 in 1994; minimum: 0 in 1997). Decreases, but not local extinction, were seen in a different site from which cattle were excluded from 1985 onwards (Kirby Canyon, maximum: 135,000 in 1992; minimum: 25,000 in 1997). Plants: Lower cover of Plantago erecta (a host plant of this butterfly) was found in sites with cattle excluded, compared to cattle-grazed sites (Silver Creek and Kirby Canyon: 4–8% cover; Coyote Ridge: 16% cover). Higher cover of total grasses (54–62% vs 25–35%), and also some non-native grasses (Lolium multiflorum: 45% vs 18–32%; Avena sp.: 18% vs 2%), was found in sites with cattle excluded. No differences were found in the cover of Vulpia microstachys, Bromus hordaceous, and B. rubens (about 2% cover for each). Methods: Postdiapause butterfly larvae were sampled in 1991–1997 in three sites with cattle excluded (two in Silver Creek, one in Kirby Canyon) and three sites with cattle (Coyote Ridge). Plants were sampled in 1996.

 

13 

A replicated, randomized, controlled study in 1993–1997 in grasslands in northern California, USA, found that plant community composition changed and the cover of herbaceous vegetation was higher in ungrazed plots, compared to cattle-grazed plots. Plants: Herbaceous plant cover was higher in ungrazed plots, compared to grazed plots, along creeks for three of six years (84–87% vs 46–59% cover). There were no differences in cover for plots by springs (data not provided). The plant community changed significantly in one plot with moderate grazing intensity but not in ungrazed plots (data reported as eigenvalues). Methods: From 1993 to 1997, three pastures in each of three areas were ungrazed, lightly grazed, or moderately grazed (three replicates of each). Cattle were allowed on grazed pasture in November and February–April each year. Plants were monitored at springs and along creeks in each pasture, each spring. Before the study, the area had been moderately grazed since 1960.

 

14 

A replicated, controlled study in 1996–1999 in rangelands in Israel (same study as (19)) found more species of plants in plots from which grazers were excluded, in two of four sites, but fewer species in one site. More plants and more plant biomass were found in plots from which grazers were excluded. Plants: More plant species were found in ungrazed plots, in two of four sites (5–12 vs 5–9 species/quadrat), but fewer were found in one site (6–13 vs 9–15). Total plant biomass was higher in ungrazed plots (10–490 vs 10–155 g/m2), and there were more plants in ungrazed plots in three of four sites (28–134 vs 28–94 plants/quadrat). There were more plants and plant species in grazed sites in less productive areas, but not in more productive areas. The abundance of common species increased with productivity in ungrazed plots, but not in grazed plots, in less productive sites. There was no difference for rare and abundant species. In more productive sites, the abundance of rare plants increased with productivity in grazed, but not ungrazed sites. There were no differences between grazed and ungrazed sites for common or abundant species. Methods: Four sites (one considerably more productive than the others) were established in 1993, each with a 10 x 10 m fenced plot to exclude sheep. Plants were surveyed in April 1996–1999, when vegetation was at a peak.

 

15 

A replicated, randomized site comparison in 1999 in blue oak savanna in the foothills of the Sierra Nevada, California, USA, found lower coverage of alien plants in ungrazed sites, compared to cattle-grazed sites. Plants: Similar numbers of alien plant species, but lower coverage of alien plants, were found in ungrazed sites, compared to cattle-grazed sites (40% vs 55% alien cover; 20 vs 26 alien species). Grazed and ungrazed sites had fewer native plants than alien plants, and fewer native plants were found where there were fewer alien plants, at one of three scales (1 m2: data reported as test statistics). Fewer perennial herbs were found on ungrazed sites, compared to cattle-grazed sites (data reported as test statistics). Methods: Grazers were excluded from five sites in Sequoia National Park at least 100 years before the study began, but they were not historically excluded from five sites on the nearby Bureau of Land Management land. Plants were sampled in 0.1 ha sites, at three scales (1 m2, 100 m2, and 1,000 m3).

 

16 

A site comparison in 2001 in grassland in southern California, USA, found more native plant species and lower cover of non-native plants in an area with cattle excluded, compared to a grazed area. Plants: More plant species were found in an area of grassland from which cattle were excluded, compared to a grazed area (24 vs 12 species), and one less non-native species was found in the ungrazed area (5 vs 6). Cover of non-native species was lower in the ungrazed area, compared to the grazed area (57% vs 75%), but the cover of native plants did not differ (49% vs 46%). Methods: In 1990, a fence was used to exclude cattle from one area of a grazed grassland. In April 2001, plants were recorded in 1,000 points on each side of the fence.

 

17 

A replicated, randomized, controlled study in 1992–2002 in grazed wetlands in northern California, USA, found more families of insects in streams in ungrazed plots, compared to grazed plots, in a ten-year experiment. Plant diversity was lower at one of two grazing intensities. In a separate three-year experiment, diversity decreased in plots from which cattle were removed, but not in grazed plots. Invertebrates: There were more families of insects in streams in ungrazed plots, compared to grazed plots (data not reported). Plants: In a ten-year experiment, there was no difference in the number of plant species, or the relative cover of native and non-native species, in ungrazed plots, compared to grazed plots (data not reported). Lower plant diversity was found after the experiment, compared to before, in ungrazed plots, compared to lightly grazed plots (but not moderately grazed, plots). Plant community composition differed between plots with or without grazers along creeks, but not at springs (data reported as ordination scores). Plant cover at the end of the experiment was higher in ungrazed plots, compared to moderately grazed plots but not lightly grazed plots (data not reported). In a separate three-year experiment, plant diversity decreased in ungrazed plots, but not in grazed plots. Methods: The ten-year experiment from 1992–2002 was established in three meadows. Within each, three watersheds were randomly assigned to one of three grazing intensities: cattle excluded, light grazing (leaving 800–1,000 pounds of residual dry matter at the end of the season), or moderate grazing (leaving 600–700 pounds). Samples were taken from both the spring and along the creek in each watershed. The three-year experiment was in 1999–2002 in marshy areas within four meadows. Two plots were established in each: one ungrazed and one with moderate grazing. Insects were surveyed every three months in one year. Plants were surveyed each June using line transects.

 

18 

A replicated, randomized, controlled study in 1995–1998 in forested pastures in central California, USA, found no difference in plant cover on stream banks and the surrounding grass between ungrazed pastures and most cattle grazing regimes. Plants: There was no difference in plant cover on stream banks and the surrounding grass between ungrazed and grazed pastures, for three of four grazing regimes (6–94% cover). There was higher plant cover in ungrazed plot, compared to grazed plots, for high intensity, dry-season grazing (72–94% vs 31–51%). Methods: One pasture in each of three streams was ungrazed and the other four were grazed moderately or intensively (reducing stubble to 2–3 and less than 2 inches, respectively) and in the dry season or the wet season (July–October and October/November–May, respectively). Plant cover was measured in June on 10 transects across the streams.

 

19 

A replicated, controlled study in 1996–1999 in shrublands in Israel (same study as (14)) found that total plant cover increased with grazer exclusion, and plant communities differed between plots with or without grazers excluded. Differences depended on the productivity of the site. Plants: Total plant abundance was higher in plots with grazers excluded, compared to grazed plots (data reported as model results). There were bigger differences in plant communities between ungrazed and grazed plots in more productive areas (data reported as Sorenson’s quantitative similarity index). Of the 36 most common annual plants, 20 showed a response to grazer exclusion: 11 species increased after grazer exclusion (only in low-productivity sites, in two of 11 species); seven species decreased (only in high-productivity sites, in one of seven species); and two species decreased in high-productivity sites but increased in low-productivity sites. Overall, more species increased than decreased in low-productivity sites (6–11 vs 1–2 species), but the opposite was true in high productivity sites (6 vs 15). Generally, large species were more abundant in ungrazed plots, compared to grazed plots (15% vs 9% relative abundance), smaller species were less abundant (54% vs 63%), and medium species showed variable responses (20% in both). These responses were more pronounced in high-productivity sites. Methods: Four 10 x 10 m plots from which sheep were excluded were established in 1993 in each of four sites, differing in topography and productivity. Vegetation samples were collected in April 1996–1999.

 

20 

A replicated, randomized, controlled study in 1991–1994 in grassland and blue oak Quercus douglasii savannas in central California, USA, found that live plant cover, native plant cover, and plant biomass were lower in areas with high numbers of ground squirrel burrows in grazed plots, but not in ungrazed plots. Mammals: The number of active ground squirrel burrows, relative to pre-experiment numbers, did not differ between ungrazed and grazed plots (60–100% vs 40–100% of pre-experiment numbers). The spatial distribution of burrows did not differ between ungrazed and grazed plots (2.6–3.4 vs 2.2–4.1 m between nearest burrows). Plants: Live plant cover, plant biomass, and native plant cover did not decrease with increasing numbers of ground squirrel burrows in ungrazed plots, but did decrease in grazed plots (3%, 60 g/m2, and 1.8% declines, respectively, for every additional burrow in a colony). Methods: Three sites, each with four plots, were established in 1991. Half of the plots were in grassland, and half were in savanna. Half had cattle excluded from them by a fence, and half were grazed from spring to summer. Three ground squirrel colonies were monitored in each plot, and vegetation was measured in a 625 cm2 plot near the centre of each, at the end of the 1992–1994 growing seasons.

 

21 

A replicated, randomized, controlled, before-and-after study in 2000–2003 in wet alpine meadows in central California, USA (same study as (22)), found fewer aquatic invertebrate species, and greater declines in the number of native plant species, in plots from which cattle were excluded, compared to grazed plots. Lower native-plant cover and higher exotic-grass cover was found in ungrazed plots, compared to grazed plots. Invertebrates: There were fewer aquatic invertebrate species in pools in ungrazed plots, compared to pools in grazed plots, in one of three years (10 vs 11–14 species). Plants: The number of native plant species in pool edges and surrounding dry land declined more in 2001–2003 in ungrazed plots, compared to grazed plots (pool edges: 1.3 fewer species vs 0.5 fewer to 1.8 more species; dry land: 0.5 fewer vs 1–1.2 more). Changes within pools did not differ between ungrazed and grazed plots. There was lower relative cover of native species in pool edges and surrounding dry land in ungrazed plots, compared to grazed plots (pool edges: 40% vs 54–72; dry land: 13 vs 18–31). There was no difference in relative cover of native species within pools between ungrazed and grazed plots. There was higher cover of exotic grasses in ungrazed and dry-season-grazed plots, compared to continuously-grazed and wet-season-grazed plots (84–86% vs 52–70%). Methods: Twenty-four plots were established in 2000, each with three pools (70–1,130 m2) and nine times more dry land than pool. Areas were either continuously or seasonally grazed by cattle (dry season: October–November; wet season: April–June), or grazers were excluded. Before the experiment, the area had been grazed for at least 100 years.

 

22 

A replicated, randomized, controlled study in rangelands in central California, USA (same study as (21)), found higher grass cover in plots from which grazers were excluded, compared to cattle-grazed plots. Plants: Grass cover was higher in ungrazed plots, compared to grazed plots (54% vs 30% cover). Methods: Thirty-six pools in 12 groups across a cattle ranch were studied, 18 of which (six groups) had fences erected around them to exclude cattle. The rest of the ranch was grazed at a density of 1 cow-calf pair/ha. Plant cover was monitored in the pools, edges, and surrounding dry land.

 

23 

A replicated, controlled study in 2000–2002 in pastures and forests on Robinson Crusoe Island, Chile, found seasonal variation in the number of species found in ungrazed plots, but not in grazed plots. Species diversity did not vary in ungrazed plots, but did in cattle-grazed plots. The cover of two of 22 species of plants varied over the course of the experiment. Plants: There was seasonal variation in the number of species found in ungrazed plots, but not grazed plots (5–8 vs 4–7 species). Diversity did not vary over time in ungrazed plots, but did in grazed plots (data reported as Shannon-Weaver-Weiner values). The cover of two of 22 species of plants varied over the experiment: Acaena argentea did not vary in cover over time in ungrazed plots, but did in grazed plots (found in 69–85% vs 51–69% of points); Conium maculatum showed variation in cover over time in ungrazed plots, but it was not found in grazed plots (found in 2–28% of points). Plant height varied over time only in ungrazed plots (22–78 vs 22–29 cm). Methods: In 2000, six 25 m2 plots were established: three in grazed pastures and three inside a cattle-exclusion fence in grazed pastures. These were then compared to six more plots in the grazed areas (three 13 m from the fence, three 44 m from the fence). Plants were monitored every three months at 121 points within each plot.

 

24 

A replicated, randomized, controlled study in 1993–2002 in wetlands in central California, USA, found higher plant diversity in plots with cattle excluded, compared to grazed plots. Herbaceous plant cover increased over time in ungrazed plots, and lightly grazed plots, but decreased in moderately grazed plots. The species composition of plant communities differed between ungrazed and grazed plots, for one of two habitats. Plants: One measure of plant diversity along creeks was higher in ungrazed plots, compared to grazed plots. Another measure along creeks was lower in ungrazed plots, compared to moderately but not lightly grazed plots. There were no differences in diversity at springs. Herbaceous plant cover increased over time in ungrazed and lightly grazed plots, but decreased in moderately grazed plots (data reported as model results). The species composition of plant communities differed between ungrazed and grazed plots for creeks, but not springs (data reported as statistical results). Plant communities showed more variability over time in ungrazed, compared to grazed plots (data presented as coefficients of variation). Methods: In 1993, three wetlands in each of three watersheds were assigned to light grazing (reducing dry matter to 250 g/m2), moderate grazing (reducing dry matter to 150 g/m2), or no grazing (grazers excluded). Plots of 2–5 ha were established at springs and creeks in each wetland. Vegetation biomass was sampled each year in 1993–1998 from three 0.0625 m2 quadrats in each plot, and plant communities were sampled with transects in 1993–2002.

 

25 

A replicated, controlled study in 1995–2001 in upland forest pastures in northeast Spain found that both shrubby and herbaceous vegetation increased more in plots from which cattle were excluded than in grazed plots. Plants: Shrub and herbaceous biomass increased in ungrazed plots, but not in grazed plots (shrub: 530 kg dry matter/ha/year; herbaceous: 220 kg). After six years, shrub and herbaceous biomass were higher in ungrazed plots, compared to grazed plots (shrub: 5,100 vs 1,200 kg dry matter/ha; herbaceous: 1,700 vs 680 kg). The number of shrubs did not differ between grazed and ungrazed plots (17–26 shrubs/transect). Herbaceous biomass contained a lower proportion of living vegetation in ungrazed plots, compared to grazed plots (23–44% vs 42–75%). Methods: In 1995, four 10 x 10 m plots were established to exclude cattle in forest pastures, which had been grazed at low intensities each March–June and October–December since 1985. Vegetation inside and outside the plots was monitored each December using transects and random points.

 

26 

A replicated site comparison in 1999–2001 in grasslands in central Spain found more species and individuals of small mammals, and higher plant biomass in plots from which cattle were excluded, compared to grazed plots. Mammals: More individuals and species of small mammals were found in plots from which cattle were excluded, compared to grazed plots (3–6 vs 0 individuals/plot; species data reported as ordination results). Three species of mammal were found: white-toothed shrews Crocidura russula (61.6% of all captures), common voles Microtus arvalis (31.9%), and wood mice Apodemus sylvaticus (6.5%). Abundances of all three species appeared to be higher in ungrazed plots, although this was not tested. Plants: Plant biomass was higher and plants were taller in plots from which cattle were excluded, compared to grazed plots, although plant cover did not differ (reported as principal component analyses) Methods: Six plots to exclude cattle were established in reforestation areas in grasslands grazed at 2–10 animals/ha. These areas were used to move livestock until the 1950s and they were reforested in 1990, but few planted trees survived. Eight live traps were placed in each of 22 trapping plots (11 inside and 11 outside cattle exclosures). Traps were set and vegetation was monitored during autumn 1999 and 2000 and summer 2000 and 2001.

 

27 

A replicated, randomized, controlled study in 2003–2005 in shrubland in north-east Israel found that plant communities in formerly grazed plots became more similar to those in ungrazed plots when cattle were excluded for two years. Plants: Plant community composition became more similar between formerly grazed and ungrazed plots over the course of the experiment (data reported as Sorensen’s quantitative similarity index). At the start of the experiment, the biomass of one of six plant functional groups differed between formerly grazed and ungrazed plots, but none differed after two years of cattle exclusion. Methods: In February 2003, fences to exclude cattle were erected around five 10 x 10 m plots in six areas: two grazed at either 0.55 or 1.1 cows/ha/year and two in areas from which cattle had been excluded for 30–40 years. Vegetation was sampled in spring of each year.

 

28 

A replicated, randomized, controlled study in 1997–2000 in grassland in South Australia found fewer native plant species in plots with grazers excluded, compared to grazed plots. There was no difference in the number of non-native species. Ungrazed plots had higher cover of two grass species in most years. Plants: Fewer native species were found in ungrazed plots, compared to grazed plots, in three of four years (1.4–2.1 vs 2.0–2.5 species/quadrat). There was no difference in the number of non-native species (2.4–4.8 species/quadrat). Cover of Austrostipa sp. (a native grass) was higher in ungrazed plots, compared to grazed plots, in all years (18–31% vs 7–16% cover). Cover of Avena barbata (a non-native grass) was higher in ungrazed plots in the last two years of the study (15% vs 6–7%) and no different from grazed plots in the first two (4–13%). Total plant cover was higher in ungrazed, compared to grazed plots (data not provided). There were no differences in the presence or absence of different species between ungrazed and grazed plots. Methods: Six 50 x 50 m plots were established in November 1997. Half were fenced to exclude grazers and half were grazed by combinations of sheep, cattle, and alpacas Vicugna pacos. Vegetation was monitored in twenty-five 1 x 1 m quadrats in each plot in November–December each year.

 

29 

A replicated, randomized, controlled study in 2002–2004 in grasslands in northern California, USA, found similar numbers of ants in plots with or without cattle excluded. Invertebrates: Similar numbers of ants were found in ungrazed plots, compared to grazed plots (data not reported). Methods: A total of eighteen 30 x 30 m plots were established in winter 2002–2003 in two sites normally grazed (one cow-calf pair/8 ha, between November and May). Twelve plots were fenced to prevent cattle grazing, of which six were also burned. Ants were surveyed with pitfall traps 14 days after burning and one year after burning.

 

30 

A replicated, randomized, controlled before-and-after study in 2003–2005 in shrubland in northern Israel found that plant diversity decreased in previously grazed plots from which cattle were excluded, but only in moderately-grazed plots, and not in heavily-grazed plots. Plants: The number of plant species decreased in plots, after grazers were excluded, at one of two grazing intensities (data reported as effective number of species). Species diversity decreased in all plots, whether or not they had been previously grazed. Methods: Areas were ungrazed or grazed at moderate or heavy stocking densities (0.55 and 1.1 cows/ha/year, respectively) for 10 years before the start of this study. In 2003, two sites were established in each of these areas (0.4–2 ha for ungrazed plots, 20–30 ha for grazed plots). Fences were erected around five 10 x 10 m plots in the grazed sites, and vegetation was monitored in these plots and also in corresponding plots in the ungrazed area.

 

31 

A controlled study in 2005–2008 in restored riparian habitat on a farm in the Central Valley, California, USA, found more plant biomass in plots without grazers, compared to plots grazed by goats and sheep. Plants: More plant biomass was found in plots without grazers, compared to plots with grazers (data reported as model results: grazing explained 21% of the variation in biomass). One-third of the identified plant species were planted during restoration (21 of 68 species), and 47 of 68 species were non-natives. Methods: Grazers were introduced to half of a streambank in 2005 (14 animals/ha), but they were excluded by a fence from the other half. Herbaceous biomass was collected in the ungrazed area and the grazed area in October 2007 and April–May 2008 (0.25 x 0.5 m plots).

 

32 

A replicated, randomized, controlled study in 2008–2010 in grasslands in central California, USA, found no difference in the cover of exotic species in plots from which cattle were excluded, compared to grazed plots. Plants: The cover of exotic species did not differ between plots from which cattle were excluded and grazed plots (data reported as model results). Methods: Ten sets of plots were established in grassland that had been grazed for decades: five in 2008 and five in 2009. Half of the plots were fenced to exclude cattle and half were left open and typically grazed in winter at approximately 0.25 cow-calf pairs/ha.

 

33 

A replicated site comparison in 2007–2008 in marshes in northeast California, USA, found that California black rails Laterallus jamaicensis were more likely to occupy ungrazed areas, compared to cattle-grazed areas, in unirrigated marshes, but were less likely to occupy ungrazed areas in irrigated marshes. Ungrazed marshes had greater cover of wetland vegetation in one of two years. Birds: The probability of marshes containing California black rails was higher for ungrazed areas, compared to grazed areas, that were not irrigated, but lower for ungrazed areas, compared to grazed areas, that were irrigated (results reported as model coefficients). Plants: The cover of wetland vegetation was higher in ungrazed areas, compared to grazed areas, in one of two years (2007: 57–65% vs 40–58% cover; 2008: 58–69%). Methods: Fourteen ungrazed marshes (fenced between 1998–2005) and 20 winter-spring grazed marshes were surveyed for rails up to twice monthly in April–August (2007) or February–October (2008, excluding September). Vegetation was sampled every month in April–July (2007) or March–August (2008). Not all marshes were surveyed for rails or vegetation each year.

 

34 

A replicated, randomized, controlled study in 2006–2010 in alpine meadows in central California, USA, found no difference in non-woody plant cover in pools in meadows from which cattle were excluded, compared to pools in grazed meadows. Plants: There was no difference in the cover of non-woody plants in pools in ungrazed meadows, compared to grazed meadows (56–80% cover). Methods: Nine meadows were studied, with cattle completely excluded from three meadows in 2006–2008 and excluded from Yosemite toad Bufo canorus breeding habitat in three meadows. The other three meadows were grazed over summer. All meadows had previously been grazed for at least a decade before the study. Plant cover was measured each summer in transects across the pools.

 

35 

A replicated, paired, site comparison in 2010–2011 in alpine meadows in central California, USA, found few differences in invertebrate communities, green plant cover or plant height between grazed and ungrazed meadows. Invertebrates: Similar numbers of invertebrate species were found in ungrazed or grazed meadows (21–34 species/sample). However, there were more species in ungrazed meadows, compared to grazed meadows, in one of four comparisons (non-ground-dwelling invertebrates sampled in mid-grazing season: data not reported). There were more individuals in ungrazed meadows, compared to grazed meadows, in mid-season, but not early season (data not reported). Three of 99 families of invertebrates had more individuals in ungrazed, compared to grazed meadows (data not reported). There were no differences in nine other measures of invertebrate communities between ungrazed and grazed meadows. Plants: Green plant cover and plant height did not differ between grazed and ungrazed meadows (54–76% cover and 7–15 cm height). Methods: Ten pairs of meadows were selected in 2010: one that had not been grazed for at least two decades, and one grazed by cattle from July–September with an average stocking density of 18.5 grazed nights/ha/year. Invertebrates and plants were sampled in July/August and September each year at four points within each meadow.

 

36 

A replicated, randomized, controlled study in 2006–2010 in alpine meadows in central California, USA, found no difference in the density of young Yosemite toads Anaxyrus canorus, the density of tadpoles, or the proportion of pools occupied by toads between plots with or without cattle excluded. Amphibians: The densities of young toads and tadpoles, and the proportion of pools that were occupied by toads, did not differ between plots with or without cattle excluded (0–12 toads/ha, 20–4,100 tadpoles/ha, 15–63% occupied pools). Methods: In 2005, 14 meadows were selected and randomly assigned to one of three treatments: unfenced (grazed, five meadows), fenced around toad breeding areas (five meadows), or fenced around the whole meadow (four meadows). Before the experiment, all meadows had been grazed been late June and September. Tadpoles were surveyed once/summer and young toads were surveyed twice/year.

 

37 

A replicated, randomized, controlled study in 2001–2005 in upland shrub pastures in northeast Spain found that shrubs and herbaceous vegetation grew faster in ungrazed plots, compared to cattle-grazed plots, resulting in greater biomass by the end of the study. Plants: Shrub biomass increased faster in ungrazed plots (2,600 vs 1,200 kg dry matter/ha/year). After five years, shrub biomass was higher in ungrazed plots (14,000 vs 6,500 kg dry matter/ha). Similar numbers of shrubs were found in ungrazed plots or grazed plots (18–45 vs 18–41 plants/transect). Herbaceous vegetation increased by 290 kg/ha/year in ungrazed plots, but did not increase in grazed plots. Herbaceous biomass was higher in ungrazed plot, in four of five years (2,100–2,800 vs 990–1,800 kg dry matter/ha). After five years, the percentage of dead herbaceous biomass was higher in ungrazed plots, compared to grazed plots (42% vs 21%). Methods: Twelve 10 x 10 m plots were established in 2001, in six shrub-dominated pastures that were grazed by cattle or sheep. Plots were fenced to exclude livestock, and vegetation was monitored with transects, quadrats, and random points in April and December each year.

 

38 

A replicated, controlled study in 2002–2007 in oak savannas in central Spain found that excluding grazers had different effects on plants at different elevations. Plant height was higher in ungrazed plots, compared to grazed plots. Plants: Fewer plant species were found in plots from which sheep were excluded, compared to plots grazed by sheep, at low elevation (6–16 vs 10–20 species/sample), but, at high elevation, more plants species were found in ungrazed plots, in two of six years (14–17 vs 13–15). Live plant cover was higher in ungrazed plots, compared to grazed plots, at high elevation (41–71% vs 34–52% cover), but it was lower at low elevation (34–95% vs 84–95%). Plant height was higher in ungrazed plots, compared to grazed plots (3–50 vs 3–34 cm). Methods: Thirty 36 m2 plots were established in 2001 at low elevation (highly productive) and high elevation (less productive), either open to grazing by sheep and rabbits, fenced to exclude sheep, or fenced to exclude both sheep and rabbits. Plants were monitored every year in April (in the low productivity site) or May–June.

 

39 

A before-and-after study in 1980–2012 on Santa Cruz Island, California, USA, found that the cover of woody vegetation increased and the cover of exotic grasses and bare ground decreased following the eradication of feral sheep from the island in 1984. Plants: Cover of woody vegetation estimated using transects increased from 1980 to 2012 (1% vs 24%), whilst bare ground decreased (40% vs 9%) and the cover of herbaceous vegetation did not change (60% vs 67%). Woody overstory plant cover (estimated from photographs) increased from 27% in 1979/1980 to 53% in 2009. Total woody vegetation cover across the island (estimated from aerial photos) increased from 26% to 77% between 1985 and 2005. Cover of non-native grasses and bare ground decreased (grasses: 68% vs 21%; bare ground: 7% vs 2%). Methods: Vegetation was monitored using transects (1980, 2012), photographs (1979/1980, 2009), and aerial photographs (1985, 2005). Before eradication, sheep had grazed the island since around 1850, at a density of approximately 2 sheep/ha (in 1980).

 

40 

A replicated site comparison in 2013 in oak woodlands in northern California, USA, found higher densities of young coast live oaks Quercus agrifolia in areas that were not grazed by cattle, compared to grazed areas. Young oaks were also larger in ungrazed, compared to grazed, areas, but there were no differences in density or size of adult trees. Plants: There were higher densities of oak seedlings and saplings in areas without cattle, compared to grazed areas (22 vs 11 trees/200 m2), but there were no differences in the density of adult trees (2 trees/200 m2). Trees were larger in ungrazed areas, compared to grazed areas (data reported as model results). Trees were less likely to have grazing damage in ungrazed areas, compared to areas with cattle, and damage was less likely to be serious (0% vs 6% of trees with at least 70% of edible biomass damaged). Methods: Areas of open oak woodland in eight ranches (four no-longer grazed, three with year-round grazing, and one with grazing from November–May) were surveyed in 2013 for oak trees of all ages using belt transects.

 

41 

A replicated, randomized, controlled study in 2008–2011 in lowland grasslands in central California, USA, found smaller increases in the number of native plant species in ungrazed plots, compared to cattle-grazed plots, but found no change in the cover of native or non-native species. Plants: The number of native species declined, or increased more slowly, in ungrazed plots, compared to grazed plots (1.8 fewer to 1.0 more species vs 1.2–2.4 more species). Change in the cover of native and non-native species did not differ between ungrazed and grazed plots (native species: 8% decline to 5% increase; data not reported for non-native species). Native, non-grass species tended to be less affected by grazing than non-native grass species (results reported as principal response curves analysis). Methods: In 2008, five experimental blocks were established, each with two 5 x 5 m plots: one that excluded cattle and one that was grazed. Plants were surveyed in 0.5 x 0.5 m quadrats each spring. Before the experiment, the area had been grazed for several decades at 0.25 animal units/ha.

 

42 

A replicated, randomized, controlled study in 2007–2013 in grasslands in central California, USA, found that native plant species were lost from plots from which cattle  were excluded, but increased in grazed plots, in one of two experiments. There were also differences in the plant community between plots with and without grazers, in one of two experiments. Plants: The number of native plant species decreased in plots from which cattle were excluded but increased in plots grazed by cattle, in one of two experiments (experiment 1: 4.5 species/year lost vs 0.5 species/year gained; experiment 2: 2.2 species/year lost vs. 0.5 species/year gained). The cover of native non-grass species increased more slowly in ungrazed plots, compared to plots grazed by cattle (data reported as log response ratios). The cover of grasses was higher in ungrazed, compared to grazed plots, in one of two experiments (experiment 1: 25–61% cover vs 8–47%; experiment 2: 17–58% for both). The diversity of native non-grass species increased more slowly in plots from which grazers were excluded, compared to plots grazed by cattle, in one of two experiments (data reported as log response ratios). Community composition varied between ungrazed and grazed plots, in one of two experiments (data reported as canonical regression coefficients). The number of native species, cover of native, non-grass species, and the diversity of measures were also less stable over time in plots from which grazers were excluded, compared to cattle-grazed plots (data reported as the ratio of average to standard deviations). Methods: Experiment 1 was established in 2007 and experiment 2 was established in 2009, both in cattle-grazed grassland. In each experiment, ten 5 x 5 m plots were established, with cattle excluded from half and allowed to graze the other half. Plants were monitored in March–April 2008–2013 in two 0.5 x 0.5 m quadrats within each plot.

 

43 

A replicated controlled study in 2012–2013 in grasslands in central California, USA, found fewer species but higher cover of native plants in plots not grazed by cattle, compared to grazed plots. Cover of invasive species and the emergence of native seedlings did not differ between grazed and ungrazed plots. Plants: Fewer species of native plants were found in plots from which cattle were excluded, compared to grazed plots (8.5–8.8 vs 10.2 species/m2). The same number of native seedlings emerged in grazed and ungrazed plots (0–1.1 seedlings/m2). Cover of native plants was higher in plots from which cattle were excluded, compared to grazed plots (72–83% vs 55–65% cover). Cover of invasive species did not differ between grazed and ungrazed plots (18–29% cover). Methods: Sixty 1 x 1 m plots were established in summer 2012: half in an area grazed at 0.25 cow-calf pairs/ha and half in an area fenced in 2012 to exclude cattle. Plant species and cover was assessed once in 2013.

 

44 

A replicated, randomized, controlled, before-and-after study in 2000–2010 in grasslands and wetlands in central California, USA (same study as (21)), found fewer species and lower cover of native plants in plots with cattle excluded, compared to grazed plots. Community composition differed, and biomass at the end of the summer was higher, in ungrazed plots, in some comparisons. Plants: Fewer native plant species were found in ungrazed plots, compared to cattle-grazed plots, in all but the first year of the experiment (6–8 vs 7–9 species/sample). Cover of native plant species was lower in ungrazed plots, compared to cattle grazed plots (32–61% vs 50–67%). This effect was stronger at pool edges than within pools or on dry land (lower native cover in four of 10 years for edges vs two of 10 years for other habitats). Ungrazed plots were more dominated by grasses in five of 10 years (0.7–4.6 vs 0.3–0.8 times more grass than non-grass cover). Ungrazed plots had higher plant biomass at the end of the grazing season, compared to grazed plots, in seven of 10 years (2,100–4,000 vs 860–2,200 kg dry mass/ha). Methods: Twenty-four plots were established in 2000, each with three pools (70–1,130 m2) and nine times more dry land than pool area. In 2000–2003, cattle were excluded from six pools, six were grazed continuously from October to June, and 12 were grazed seasonally (either October–January or April–June). In 2003, the seasonal grazing experiment was stopped and only ungrazed and continuously grazed plots were continued. Plants were monitored in April–May each year within pools, at pool edges, and on dry land. The area had been grazed for at least 100 years before the start of the experiment.

 

45 

A site comparison in 2014 in an alpine meadow in southern California, USA, found more and larger willow Salix sp. trees in a plot from which cattle were excluded, compared to a grazed area. Plants: Streamside vegetation within a plot from which cattle were excluded had more willows than the surrounding grazed area (980 trees in 1,200 m vs 75 in 900 m) and had a lower cover of sedges Carex sp. (data not reported). Willows within the ungrazed plot were larger than those in the grazed area (average height: 48 vs 25 cm; maximum height: 300 vs 170 cm). Methods: Cattle were excluded from a 1,200 m section of stream in 1991 but were allowed to graze on the remaining 900 m. Vegetation within 2 m of the stream was surveyed in 2014.

 

Referenced papers

Please cite as:

Shackelford, G. E., Kelsey, R., Robertson, R. J., Williams, D. R. & Dicks, L. V. (2017) Sustainable Agriculture in California and Mediterranean Climates: Evidence for the effects of selected interventions. Synopses of Conservation Evidence Series. University of Cambridge, Cambridge, UK.