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: Other biodiversity: Use fewer grazers Mediterranean Farmland

Key messages

Read our guidance on Key messages before continuing

Amphibians (0 studies)

Birds (0 studies)

Invertebrates (1 study): One replicated, randomized, controlled study in wet grasslands in the USA found more families of insects in streams in areas grazed by cattle at lower, compared to higher, intensities.

Mammals (0 studies)

Plants (11 studies)

  • Abundance (11 studies): Six studies (four replicated, randomized, and controlled) in grasslands or wood pasture in the USA, Chile, and Israel found higher cover of some species of plants, herbaceous plants, or native plants in areas grazed by cattle or sheep at lower, compared to higher, intensities. One controlled study in forest in Israel found higher cover of woody vegetation in areas with lower grazing intensity. Four of these studies also found lower cover or biomass of some groups of plants in sites with lower grazing intensity. Four studies in grasslands in the USA and Israel found no effect of grazing intensity on biomass, cover, or abundance of plants.
  • Diversity (6 studies): Three replicated, randomized, controlled studies in grasslands and wet grasslands in the USA and Israel found no differences in plant diversity between sites with different cattle-grazing intensities, in some or all comparisons. One of these also found higher diversity in some comparisons and lower diversity in others. One replicated, randomized, controlled study in wet grasslands in the USA found that plant community composition differed in sites with different cattle-grazing intensities, in some comparisons. Two replicated, randomized, controlled studies in grasslands and wet grasslands in Israel and the USA found no differences in the number of plant species between sites with different cattle grazing intensities, in some or all comparisons. One of these studies also found more species in some comparisons and fewer species in others. One controlled study in wood pasture in Chile found fewer native species and more non-native species in paddocks with lower sheep-grazing intensities.
  • Survival (3 studies): Three controlled studies (two replicated and randomized) in grasslands in the USA and forests in Israel found no difference in native grass, tree, or shrub survival in areas grazed by cows at lower, compared to higher, intensities.

Reptiles (0 studies)

Implementation options (0 studies)

Supporting evidence from individual studies

1 

A controlled study in 1979–1984 in improved and wooded grasslands in northern California, USA, found that the cover of different plant species varied with sheep-grazing intensity, and that the effects differed between improved and wooded grasslands. Plants: On improved grasslands, cover of one of five species was higher on plots with lower grazing intensity, whilst cover of two species was lower. Cover of two species either peaked at intermediate grazing intensities or did not vary. In wooded grasslands, cover of three of eight species was higher on plots grazed at lower intensities, whilst cover of five was either highest or lowest at intermediate intensities, or did not vary. In improved grassland, total herbaceous plant biomass was highest under intermediate grazing and lowest under high intensity (data not provided). In wooded grasslands, total herbaceous plant biomass did not vary with grazing intensity (data not provided). Methods: Improved grassland was seeded with subterranean clover and fertilized with sulphur (12 kg/ha) and triple super phosphate (11.3 kg P/ha). One plot in each habitat was grazed at each of low, medium, and high intensities (0.6, 1.8 and 3.1 sheep/ha for woodland and 5.3, 8.0 and 10.0 for improved grassland, respectively). Vegetation was monitored in April–May each year using point-step transects and clipping vegetation from ten 0.9 m2 plots in each pasture.

 

2 

A replicated, randomized, controlled study in 1989–1991 in grasslands in north-central California, USA, found that blue oak Quercus douglasii seedlings survived at similar rates but had less damage from cattle when planted in plots grazed by cattle at lower densities. Plants: Survival of seedlings after 15 months did not differ between plots grazed at low, medium, or high densities (7–50% survival). Fewer seedlings were damaged by browsing and trampling as grazer densities decreased (12–57%, 38–74%, and 52–93% of seedlings damaged in low, medium, and high density plots, respectively). Methods: In December 1989, oak seedlings were planted in three pastures, each containing nine plots grazed for a week each in winter, spring, or summer with low, medium and high densities of cattle (2.5, 7.5, and 15 animals/ha, respectively). 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.

 

3 

A replicated, randomized controlled study in 1993–1997 in grasslands in northern California, USA, found no difference in the cover of herbaceous vegetation between plots that were lightly or moderately grazed by cattle, but cover was more stable in lightly grazed plots. Plants: Similar cover of herbaceous vegetation was found in lightly-grazed and moderately-grazed plots, but cover was more stable in lightly-grazed plots (along creeks: 46–47% vs 35–80% cover; by springs: data not provided). Methods: In November and February–April 1993–1997, three pastures in each of three areas had either light grazing or moderate grazing (three replicates of each). In spring, plants were monitored at springs and alongside creeks in each pasture. The area had been grazed at moderate intensity since 1960.

 

4 

A controlled study in 1984–1991 in grazed broadleaf forests in northern Israel found higher cover of woody vegetation, taller trees and less herbaceous vegetation in an area moderately, compared to heavily, grazed by cattle. Plants: Woody vegetation did not decline over time in a moderately grazed area, but did in a heavily grazed area (change from 48% to 55% cover vs 49% to 41%). At the end of the study, regrowing Kermes oak Quercus calliprinos trees were taller in the moderately, compared to the heavily, grazed area (1.1 vs 0.3 m). There were no differences at the beginning of the study (0.6 m for both). Herbaceous vegetation cover and biomass increased in both areas, but increased by less in the moderately, compared to the heavily, grazed area (cover: increase from 13% to 24% vs 20% to 52%; biomass: increase from 28 to 56 g/m2 vs 17 to 218 g/m2). There was no difference in the percentage of shrubs that were alive between the areas (data not provided). Methods: In 1984, an area of woodland (mechanically cleared, treated with herbicide, and grazed for two years) was split into two areas. One was grazed moderately and one was grazed heavily (0.30 and 0.54 cows/ha, respectively). Animals were removed for December–March and May–November each year. Vegetation was monitored each April in 20–30 25 x 25 cm quadrats in each major habitat.

 

5 

A replicated, randomized, controlled study in 1993–1997 in grassland in northeast Israel (same study as (12)) found few differences in the number of plant species between plots that were moderately or heavily grazed by cattle. Plants: In early-grazed plots, more plant species were found with moderate grazing, compared to heavy grazing, in three of four years (46–49 vs 36–44 species/plot). In late-grazed plots, however, fewer plant species were found with moderate grazing, in three of four years (28–42 vs 41–50 species/plot). Similar numbers of species were found in moderately-grazed and heavily-grazed plots, with continuous grazing (49–66 species/plot). Cover of tall annual grasses was higher in moderately, compared to heavily, grazed plots, for three of four years, and two of three grazing timings (12–40% vs 5–21% cover). Cover of annual thistles was lower in moderately, compared to heavily, grazed plots for two or three out of four years, depending on the timing of grazing (4–9% vs 6–18% cover). Cover of crucifers was lower with moderate, compared to heavy, grazing for two to four years, depending on timing of grazing (1–14% vs 10–20%). With early season grazing, crucifer cover was higher with moderate grazing for one year (16% vs 8%). Seven other functional groups did not vary between plots of different grazing intensities. Methods: In 1993 eight plots were established with half grazed continuously (January–October) and half grazed seasonally. In seasonal plots, half the plot was grazed early (January–April/May) and half late (April/May–October). In addition, half the plots were subject to moderate grazing (0.55 or 1.1 cow-calf pairs/ha for continuous and seasonal respectively) and half to heavy (1.1 or 2.2 cow-calf pairs/ha). Plants were surveyed every spring every two steps along permanent transects.

 

6 

A replicated, randomized, controlled study in 1992–2002 in grazed wetlands in northern California, USA, found more families of aquatic insects in streams in lightly, compared to moderately, grazed plots. There were no differences in the number of plant species or the relative cover of native and non-native species, although diversity and community composition in one of two habitats did differ between grazing intensities and total plant cover was lower in moderately grazed plots. Invertebrates: There were more families of insects in streams in lightly grazed plots, compared to moderately grazed plots (data not provided). Plants: There were no differences in the number of plant species or the relative cover of native and non-native species (data not reported). Diversity relative to before the experiment was higher in lightly grazed plots, compared to moderately grazed plots. Plant community composition varied between lightly and moderately grazed plots along creeks, but not at springs, in wetlands (data reported as ordination scores). Plant cover at the end of the experiment was higher in lightly grazed, compared to moderately grazed, plots (data not provided). Methods: Three meadows were studied, with three watersheds in each randomly assigned to a grazing intensity: one with cattle excluded, one with light grazing (leaving 800–1,000 pounds of residual dry matter at the end of the season) and one with moderate grazing (leaving 600–700 pounds). Samples were taken from both the spring and along the creek in each watershed. Insects were surveyed every three months in one year. Plants were surveyed each June using line transects.

 

7 

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 surrounding grass between pastures moderately and intensively grazed by cattle. Plants: There was no difference in plant cover on stream banks and the surrounding grass between moderately and intensively grazed (31–84% cover). Methods: Two pastures in each of three streams were assigned to either moderate or intensive grazing (reducing stubble to 2–3 and less than 2 inches, respectively). Half of each were grazed in the dry season, and half in the wet season (July–October and October/November–May, respectively). Plant cover was measured in June on 10 transects running across the streams.

 

8 

A replicated, controlled study in 1998–2000 in north-central California, USA, found no differences in purple needlegrass Nassella pulchra survival or density in lightly, compared to heavily, grazed areas. In one of three years, plants in lightly grazed plots had more reproductive stems and were taller than those in heavily grazed plots. Plants: Needlegrass mortality and density did not vary between grazed and ungrazed plots (data reported as model results). Plants in lightly grazed plots had more reproductive stems, compared to those in heavily grazed plots in one of three years (4 vs 2). Plants were also taller in lightly grazed, compared to heavily grazed, plots in one of three years (data not provided). Plant stem diameter did not vary between lightly and heavily grazed plots (data not provided). Methods: Twenty 20 x 20 m plots were rotationally grazed from January (in 1998) or December (1999, 2000) until May at a stocking density of 0.75 animal units/ha. Plots were grazed until 25% (lightly grazed) or 50% (heavily grazed) of plant biomass was removed and then rested for 35 days. Thirty individual plants were measured each year and plant density was estimated using 3–5, 1 x 1 m quadrats in each plot.

 

9 

A controlled study in 1976–1983 in wood pasture (Espinal) in central Chile found fewer species of native plants, more species of non-native plants, more grasses, fewer composites, and more plant biomass in paddocks with lower stocking rates of sheep. Plants had different traits with different stocking rates. Plants: Fewer native species were found in paddocks with lower stocking rates, in three of four years (e.g., with 1 vs 3.5 sheep/ha, in the 8th year: 2.5 vs 8 native species), and more non-native species were found in one of four years (e.g., with 1 vs 3.5 sheep/ha, in the 6th year: 9.5 vs 7.5 non-native species). In two of four years, the highest numbers of non-native species were found with intermediate stocking rates (e.g., in the 3rd year, with 1 vs 2.5 sheep/ha: 11 vs 7.5 non-native species). Overall, fewer plant species (native and non-native) were found in paddocks with lower stocking rates (e.g., with 1 vs 3.5 sheep/ha, in the 8th year: 12 vs 15 species). More grasses, and fewer plants in the daisy family (composites), were found in paddocks with lower stocking rates (e.g., with 1 vs 3.5 sheep/ha, in the 8th year: 75% vs 25% relative frequency, for grasses; 10% vs 65% for composites), but similar numbers of legumes were found at different stocking rates (1–3%). More plant biomass was found in paddocks with lower stocking rates (e.g., with 1 vs 3.5 sheep/ha, in the third season: 850 vs 150 kg dry biomass/ha). Lower stocking rates were associated with taller, more palatable plants with fibrous roots and animal-dispersed seeds, whereas high stocking rates were associated with shorter plants that grow in rosettes and produce fewer wind-dispersed seeds (data reported as locations of morpho-functional traits in ordination space). Methods: The study area (32 ha) was grazed with 1 sheep/ha for at least 20 years before the study began. In 1976, seven paddocks were established with fences (2.5–10 ha/paddock, 10 sheep/paddock), each with a different stocking rate (1, 1.5, 2, 2.5, 3, 3.5, or 4 sheep/ha). In paddocks with 1–3.5 sheep/ha, plants were sampled in spring (October–November, five 4 m transects/plot), and plant biomass was measured in exclusion cages at the end of the growing season (one 1 m2 cage/paddock from May to December), in 1978, 1979, 1981, and 1983.

 

10 

A replicated, randomized controlled study in 1993–2002 in wetlands in central California, USA, found lower plant diversity in one of two habitats in lightly, compared to moderately, grazed plots. Herbaceous plant cover increased in lightly, but decreased in moderately, grazed plots. Species composition of plant communities differed, but variability did not, between lightly and moderately grazed plots for one of two habitats. Plants: Two measures of plant diversity were lower in lightly, compared to moderately, grazed plots alongside creeks, but not at springs (data reported as model results). Herbaceous plant cover increased over time in lightly grazed plots, but decreased in moderately grazed plots (data reported as model results). The species composition of plant communities varied between lightly and moderately grazed plots (data reported as ordination results). Plant community variability did not differ between lightly and moderately grazed plots (data presented as coefficients of variation). Methods: In 1993, three wetlands in each of three watersheds were assigned to either light grazing (reducing dry matter to 250 g/m2) or moderate grazing (reducing dry matter to 150 g/m2). 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 sampled with transects in 1993–2002.

 

11 

A replicated, randomized, controlled study in 2008–2010 in grasslands in northern California, USA, found that plant diversity did not vary between plots grazed at different intensities, but that cover of native species was higher in plots grazed at lower intensities, compared to those grazed at higher intensities, for one of three plant assemblages. Plants: Plant diversity did not differ between plots with different grazing intensities (results reported as Shannon diversity). Cover of native species was higher in plots with medium, rather than high, grazing intensity in plots sown with native perennial grasses (17% vs 22% cover), but there was no difference in plots sown with two types of non-native assemblages (2–3%). Methods: In 2007 four experimental blocks were established across two pastures. Each block was split into three areas, sown with one of three vegetation types: native perennial grasses, non-native annual forage grasses, or a non-native, non-edible annual weed. These were then divided into six plots, which were subjected to one of three treatments, each replicated twice: no manipulation, mowing and trampling by cattle to simulate medium cattle grazing, or mowing and trampling by cattle to simulate heavy cattle grazing. Plants were surveyed in a 1 m2 quadrat in May 2008–2010.

 

12 

A replicated, randomized, controlled study replicated study in 1993–2012 in grassland in north-eastern Israel (same study as (5)) found no difference in plant diversity in plots grazed by cattle at moderate, compared to heavy, intensities. The cover of three of five plant functional groups varied between moderately and heavily grazed plots. Plants: Plant diversity did not differ between moderately and heavily grazed plots (data not reported). Cover of tall annual grasses was higher in moderately, compared to heavily, grazed plots for two out three grazing timings (7–52% vs 2–23% cover, no difference for early-grazed plots). Cover of short grasses was lower in moderately, compared to heavily, grazed plots (1–30% vs 3–41%), with the biggest differences in continuously, rather than seasonally, grazed plots. Cover of tall perennial grasses was lower under moderately, compared to heavily, grazed plots for one of three grazing timings, but this difference had gone by the end of the study. There were no differences in cover of perennial or annual non-grass plants between grazing intensities. Methods: In 1993 eight plots were established with half grazed continuously (January–October) and half grazed seasonally. In seasonal plots, half the plot was grazed early (January–April/May) and half late (April/May–October). In addition, half the plots were subject to moderate grazing (0.55 or 1.1 cow-calf pairs/ha for continuous and seasonal respectively) and half to heavy (1.1 or 2.2 cow-calf pairs/ha). Plants were surveyed every spring every two steps along permanent transects.

 

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.