Exclude or remove livestock from historically grazed brackish/salt marshes

How is the evidence assessed?
  • Effectiveness
    58%
  • Certainty
    60%
  • Harms
    10%

Study locations

Key messages

VEGETATION COMMUNITY

  • Overall extent (1 study): One controlled study of a salt marsh in Germany reported that in a plot fenced to exclude cattle for eight years, the total vegetated area was greater than in a plot that remained grazed.
  • Community types (1 study): One site comparison study of brackish and salt marshes in Germany reported that reducing (or stopping) grazing affected the nature of transitions between vegetation types over time, but that the precise effect varied with environmental conditions.
  • Community composition (5 studies): Three paired studies (two also replicated and controlled) in brackish/salt marshes in France, Argentina and the Netherlands reported that the effect of excluding livestock for 5–30 years on the overall plant community composition depended on plot elevation/flooding regime. In one of these studies, the effect of livestock exclusion was not separated from the effect of general legal protection. Two studies in one salt marsh in Denmark reported that excluding livestock had little effect on the identity of plant species in the community after six years.
  • Overall richness/diversity (6 studies): Two studies (one controlled, one before-and-after) in one salt marsh in Denmark reported that excluding sheep and cattle for 6–7 years had no effect on overall plant species richness. One replicated, paired, controlled study in a salt marsh in the Netherlands reported that plots fenced to exclude cattle for seven years had lower plant species richness than areas that remained grazed. Two controlled studies (one also replicated and paired) in salt marshes in Germany found that the effect of removing sheep on overall plant species richness depended on the scale of measurement and the grazing intensity used for comparison – with inconsistent results across these studies even for similar scales and intensities. One paired, site comparison study of salt marshes in Argentina found that the effect of excluding cattle (along with legal protection) increased plant species richness at lower elevations, but did not significantly affect plant diversity at any elevation.

VEGETATION ABUNDANCE

  • Overall abundance (4 studies): Three studies (two controlled, one before-and-after) in salt marshes in the UK and Denmark reported that excluding livestock for 2–6 years maintained or increased overall vegetation abundance (although in one study, only by a small amount). One controlled study in a salt marsh in Germany found that a paddock left ungrazed for 16–18 years had greater overall vegetation cover than lightly or heavily grazed paddocks, but lower cover than a moderately grazed paddock.
  • Individual species abundance (11 studies): Eleven studies quantified the effect of this action on the abundance of individual plant species. For example, five studies (four controlled, one before-and-after) on salt marshes in the UK, Denmark, Germany and the Netherlands reported that excluding livestock for 2–8 years reduced (or prevented increases in) cover of saltmarsh grass Puccinellia maritima. However, two controlled studies (one also replicated and paired) on salt marshes in Denmark and Sweden reported greater saltmarsh grass cover in areas fenced to exclude livestock for 1–6 years than in areas that remained grazed. Four studies (three controlled, one before-and-after) on salt marshes in Denmark and Germany reported that excluding or removing livestock for 4–16 years increased cover of sea purslane Halimione portulacoides.

VEGETATION STRUCTURE

  • Height (5 studies): Five controlled studies (two also replicated and paired) in salt marshes in Sweden and Germany, and brackish wet grassland in the UK, found that ungrazed plots (livestock excluded or removed) contained taller vegetation than plots that remained grazed. Vegetation was surveyed after one month, 1–8 years or 16–22 years.

About key messages

Key messages provide a descriptive index to studies we have found that test this intervention.

Studies are not directly comparable or of equal value. When making decisions based on this evidence, you should consider factors such as study size, study design, reported metrics and relevance of the study to your situation, rather than simply counting the number of studies that support a particular interpretation.

Supporting evidence from individual studies

  1. A replicated, controlled study in 1955–1957 in an estuarine salt marsh in England, UK (Ranwell 1961) reported that excluding livestock maintained greater overall vegetation biomass than continued grazing, but had mixed effects on the abundance of dominant plant species. Unless specified, statistical significance was not assessed. At the start of the experiment, total above-ground vegetation biomass was 8,061 g/m2. After two years, this had declined to 7,118 g/m2 in exclusion plots, vs 5,633 g/m2 in grazed plots. Over two years, saltbush Atriplex hastata biomass declined less in exclusion plots (by 70%) than in grazed plots (by 277%). In contrast, cordgrass Spartina sp. biomass declined more in exclusion plots (by 97%) than in grazed plots (by 67%) and saltmarsh grass Puccinellia maritima biomass increased less in exclusion plots (by 80%) than in grazed plots (by 99%). Changes in cover were typically similar in both exclusion and grazed plots. Exceptionally, saltmarsh grass cover did not significantly change in three of four exclusion plots but significantly increased in four of four grazed plots (data not reported). Methods: In summer 1955, eight 9 x 13 m plots were established in a historically grazed salt marsh. Four plots were fenced to exclude sheep. Four plots were grazed by sheep during summer (average 24 sheep days/plot/year). Vegetation was surveyed in early June at the start of the experiment (1955) and over the two following years (1956–1957). Biomass was dried before weighing.

    Study and other actions tested
  2. A replicated, controlled, site comparison study in 1972–1974 in two brackish/salt marshes in Georgia, USA (Reimold et al. 1975) reported that plots fenced to exclude livestock had similar patterns of species dominance to grazed areas, but typically contained more vegetation biomass. Statistical significance was not assessed. Over the two years following intervention, smooth cordgrass Spartina alterniflora was dominant in all ten sampled months in both exclosures (48–73% of biomass) and grazed areas (53–66% of biomass). Saltgrass Distichlis spicata and pickleweed Salicornia virginica were subdominants. In 24 of 28 comparisons, the total above-ground biomass of these three species was greater in exclosures (28–362 g/m2) than grazed areas (15–277 g/m2). For the grasses, data were also reported for live and dead biomass separately, and showed similar patterns (see original paper). A nearby natural marsh was dominated by smooth cordgrass in 7 of 10 sampled months (51–69% of biomass) but pickleweed in the other three (54–65% of biomass). In every month with a fair comparison to the natural marsh, exclosures contained less biomass of both smooth cordgrass (exclosures: 82–345 g/m2; natural: 184–439 g/m2) and pickleweed (exclosures: 38–99 g/m2; natural: 194–490 g/m2). Methods: In May 1972, electric fences were installed to exclude livestock from five 200-m2 areas of a coastal brackish/salt marsh (salinity 5–35 ppt). The rest of this marsh remained “heavily” grazed. The study does not report details of the livestock or grazing intensity. Above-ground vegetation was sampled (cut, dried and weighed) approximately monthly over the following two years: from the exclosures, the grazed area and a nearby natural (never-grazed) marsh.

    Study and other actions tested
  3. A controlled study in 1972–1978 in a salt marsh in Denmark (Jensen 1985) reported that an area fenced to exclude livestock had identical plant species richness to an area that remained grazed, but greater vegetation cover. Statistical significance was not assessed. After approximately six years, the same seven plant species were present in the exclusion and grazed areas. However, six of these species consistently had greater cover in the exclusion area – including saltmarsh grass Puccinellia maritima (exclusion: 84–92%; grazed: 72–82%) and sea purslane Halimione portulacoides (exclusion: 8–17%; grazed: <1%). Accordingly, both sampling plots within the exclusion area had greater overall vegetation cover than both sampling plots within the grazed area. This was true for the sum of the cover of each species (exclusion: 130–145%; grazed: 81–89%) and for cover as the inverse of bare ground (exclusion: 95–98%; grazed: 73–83%). Methods: In spring 1972, an area of historically grazed coastal salt marsh was fenced to exclude livestock. The surrounding marsh remained grazed (by at least 0.5 sheep/ha and 0.5 cattle/ha, May–October). In August 1978, the cover of every plant species and bare ground were recorded in two plots in the fenced and grazed areas (50 point quadrats with 10 pins/plot). This study used the same marsh as [4], but a different experimental set-up.

    Study and other actions tested
  4. A before-and-after study in 1971–1978 in a salt marsh in Denmark (Jensen 1985) reported that after installing fences to exclude livestock, plant species richness was stable but there were small changes in vegetation composition and cover. Statistical significance was not assessed. The study plot contained nine plant species both before and approximately six years after livestock exclusion. Eight of the species were the same (one species went extinct and one new species colonized). Overall vegetation cover increased slightly. This was true for the sum of the cover of each species (before exclusion: 176%; six years after: 180%) and for cover as the inverse of bare ground (before exclusion: 92%; six years after: 98%). Changes in cover of individual species included increases for red fescue Festuca rubra (before: 55%; after: 72%) and sea purslane Halimione portulacoides (before: 22%; after: 61%) and a decrease for saltmarsh grass Puccinellia maritima (before: 34%; after: 3%). Methods: In spring 1972, a 40 x 60 m area of coastal salt marsh was fenced to exclude livestock. Previously, the marsh had been grazed by 0.5 sheep and 0.5 cattle/ha, May–October. The cover of every plant species and bare ground were surveyed in a permanent plot, using point quadrats, before (August 1971) and after (August 1978) exclusion. This study used the same marsh as [3], but a different experimental set-up.

    Study and other actions tested
  5. A replicated, paired, controlled study in 1979–1983 on a coastal salt marsh in Sweden (Pehrsson 1988) found that plots from which cattle were excluded contained taller grasses than plots open to grazing. This was true in three of three years for creeping bentgrass Agrostis stolonifera (exclusion: 15–20 cm; grazed: 6–7 cm) and saltmarsh grass Puccinellia maritima (exclusion: 10–21 cm; grazed: 5–6 cm). The study also noted broadly similar changes in cover of these species under both treatments over four years, but that cover was more often higher in exclusion than grazed plots (statistical significance not assessed; data not clearly reported). Methods: In autumn 1979, five pairs of 25-m2 plots were established on a grazed salt marsh (0.1–0.9 cattle/ha). One plot in each pair was fenced to exclude cattle. Geese were also excluded from two of these plots from 1981. Vegetation was surveyed using point quadrats each autumn between 1979 and 1983 (including height measurements for 12–348 plants/species/treatment/year, between 1980 and 1982).

    Study and other actions tested
  6. A controlled study in 1980–1988 on a salt marsh in northern Germany (Andresen 1990) reported that a plot fenced to exclude livestock had greater overall vegetation coverage, different cover of some individual plant species, and taller vegetation than a plot that remained intensely grazed. Statistical significance was not assessed. After eight years, the total vegetated area was greater in the exclusion plot than the grazed plot, especially closer to the sea (data reported as maps). In the exclusion plot, vegetation stands further from the sea became dominated by couch grass Elymus pycnanthus at the expense of saltmarsh grass Puccinellia maritima, perennial ryegrass Lolium perenne and red fescue Festuca rubra (data reported as frequency classes). The exclusion plot also contained taller vegetation (understory grasses: 7–32 cm; canopy: 41–87 cm) than the grazed plot (understory grasses: 3–14 cm; canopy: 0–55 cm). Methods: In 1980, a 10-ha plot in a coastal salt marsh was fenced to exclude cattle. Another plot in the marsh remained heavily grazed (2 cattle/ha). Vegetation stands were mapped in 1988. Vegetation was surveyed in more detail along one 750-m transect/plot, at 50 m intervals. Cover was assessed using a point quadrat (50 points/interval).

    Study and other actions tested
  7. A controlled study in 1988–1992 in a historically grazed salt marsh in northern Germany (Kiehl et al. 1996) reported that a plot where sheep grazing was stopped developed different cover of key plant species and taller vegetation than a plot that remained intensely grazed. Statistical significance was not assessed. After four years, the ungrazed plot had only 49% cover of the dominant saltmarsh grass Puccinellia maritima (vs grazed: 71%) and 0% cover of glasswort Salicornia europaea (vs grazed: 24%). Meanwhile, the ungrazed plot had 18% cover of red fescue Festuca rubra (vs grazed: 0%) and 18–21% cover of each of the perennial herbs sea aster Aster tripolium and sea purslane Halimione portulacoides (vs grazed: 0%). The perennial herbs also occurred in a greater proportion of quadrats across the ungrazed plot (see original paper for data). Finally, the ungrazed plots had taller vegetation on average (18 cm) than the grazed plot (8 cm). Methods: In 1988, sheep were excluded from one section of a coastal salt marsh. Another section of the marsh remained grazed by sheep (10 sheep/ha, April–October). The marsh was intertidal (flooded 80–200 times/year) but had a dense artificial drainage system. The cover of each plant species was surveyed annually between 1989 and 1992 in permanent quadrats. Perennial herb frequency and overall vegetation height were recorded in 1992, in quadrats or along transects respectively.

    Study and other actions tested
  8. A replicated, paired, controlled, before-and-after study in 1989–1994 of eighteen historically grazed brackish marshes in southern France (Mesléard et al. 1999) reported that the effects of excluding livestock on plant community composition, abundance and species richness depended on the flooding regime. Unless specified, statistical significance was not assessed. Under all three flooding regimes, the overall plant community composition in exclusion and grazed marshes diverged over five years. However, the speed and direction of the changes depended on the flooding regime (data reported as graphical analyses). For example, under two artificial flooding regimes, exclusion significantly increased the final cover of sea club rush Bolboschoenus maritimus (exclusion: 31–33%; grazed: 11–12%) and common reed Phragmites australis (exclusion: 12–16%; grazed: <1%). Under an unmanaged flooding regime, exclusion increased cover of the grass Aeluropus littoralis (exclusion: 16%; grazed: 0%). After five years, total plant species richness was higher in exclusion than grazed marshes under both artificial flooding regimes (exclusion: 4 species/0.25 m2; grazed: 5–6 species/0.25 m2) but lower in exclusion than grazed marshes under the unmanaged flooding regime (exclusion: 7 species/0.25 m2; grazed: 5 species/0.25 m2). Methods: The study used two sets of nine inland brackish marshes (former rice fields, but grazed since 1976 when cultivation stopped). In November 1989, one set was fenced to exclude livestock. The other set remained grazed (approximately 2 cattle and 1 horse/ha, April–November). Three of the nine 1-ha marshes within each set received each flooding regime: artificial winter flooding, artificial summer flooding, or year-round unmanaged flooding. Vegetation was surveyed every six months from early November 1989 to early November 1994 (nine 0.5 x 0.5 m quadrats/field/survey).

    Study and other actions tested
  9. A replicated, paired, controlled study in 1991–1998 in a salt marsh in the Netherlands (Esselink et al. 2002) reported that plots fenced to exclude cattle contained fewer plant species than grazed areas, and developed cover of different plant species. Statistical significance was not assessed. After seven years, plots from which cattle had been excluded had lower plant species richness (8–9 species/100 m2) than areas that remained grazed (13–14 species/100 m2). Exclusion plots were dominated or co-dominated by a mix of species, including couch grass Elymus repens (7–90% cover/plot), creeping bent Agrostis stolonifera (1–40% cover/plot) and saltmarsh grass Puccinellia maritima (0–50% cover/plot). In contrast, grazed plots were all dominated by saltmarsh grass (50–80% cover/plot). Couch grass and creeping bent showed particularly strong responses to the seven years of exclusion: cover of these species was 0–1% in the spring immediately following exclusion, and never more than 2% in grazed plots. Methods: In 1991, two 40 x 40 m plots within a grazed salt marsh were fenced and cattle were successfully excluded. One plot was closer to the sea and one closer to the land. Between 1991 and 1998, vegetation was surveyed in 10 x 10 m quadrats: two within each plot and three in the grazed marsh (38–81 animal days/ha/year) around each plot.

    Study and other actions tested
  10. A replicated, controlled, before-and-after study in 1997 of eight areas of a brackish wet grassland on the coast of England, UK (Hart et al. 2002) found that plots left ungrazed contained taller vegetation than plots grazed by cattle. After approximately one month, the vegetation was taller in ungrazed plots (6.1–10.5 cm) than in grazed plots (4.2–5.0 cm). Before intervention, vegetation height did not significantly differ between plots destined for each treatment (ungrazed: 3.6 cm; grazed: 3.2 cm). Methods: The study used eight plots on a coastal, brackish, wet grassland that had been grazed by cattle in 1996. Four plots were left ungrazed in 1997, whilst four plots were lightly grazed by cattle (0.2–0.5 livestock units/ha) from mid-April. The overall height of the grassy vegetation was measured in 1997 before grazing began (early April) and approximately one month after (mid–late May). There were 20–40 survey points/plot.

    Study and other actions tested
  11. A controlled study in 1991–2009 on a salt marsh in northern Germany (Rickert et al. 2012) found that a paddock from which sheep had been removed contained taller vegetation than paddocks which remained grazed, and typically had higher vegetation cover and plant species richness. After 16–18 years, the vegetation canopy was taller in an ungrazed paddock (20 cm) than in all grazed paddocks (6–14 cm). In two of three comparisons, overall vegetation cover was greater in an ungrazed paddock (87%) than in grazed paddocks (lightly grazed: 86%; heavily grazed: 82%). In the other comparison, cover was lower in the ungrazed paddock (vs moderately grazed: 89%). In two of three comparisons, total plant species richness was greater in an ungrazed paddock (10.1 species/m2) than in grazed paddocks (moderately grazed: 9.9; heavily grazed: 6.7 species/m2). In the other comparison, richness was lower in the ungrazed paddock (vs lightly grazed: 12.2 species/m2). The study also reported cover of the dominant species in each paddock. For example, sea purslane Atriplex portulacoides was the most abundant species in the ungrazed paddock (27% cover) but not in grazed paddocks (<7–19% cover). Methods: The study used four paddocks on a coastal salt marsh (historically heavily grazed by sheep). From 1991, livestock were removed from one paddock. The other paddocks were grazed each summer: lightly (1–2 sheep/ha), moderately (3–4 sheep/ha), or heavily (10 sheep/ha). Vegetation was surveyed every three weeks in summer 2007–2009, in a total of thirty 1-m2 quadrats/paddock/year. All quadrats were at a similar elevation (±20 cm). The paddocks in this study were also used in [12].

    Study and other actions tested
  12. A replicated, paired, controlled study in 1988–2010 on three salt marshes in northern Germany (van Klink et al. 2013) found that paddocks from which sheep had been removed contained taller vegetation than paddocks which remained grazed, but a similar number of plant species. After 19–22 years, the vegetation canopy was taller in ungrazed paddocks (25 cm) than in all grazed paddocks (5–19 cm). At a large scale, plant species richness did not significantly differ between ungrazed paddocks (10.3 species/1.1 m2) and all grazed paddocks (13.3–14.0 species/1.1 m2). However, at a smaller scale, ungrazed paddocks contained fewer plant species (4.1 species/0.33 m2) than heavily grazed paddocks (7.3 species/0.33 m2) or short vegetation patches in moderately grazed paddocks (5.8 species/0.33 m2) – but a similar number of plant species to tall vegetation patches in moderately grazed paddocks (5.3 species/0.33 m2). The study also reported cover of the dominant species in each paddock (see original paper for data). Methods: The study used nine 11–15 ha paddocks: three on each of three coastal salt marshes (historically heavily grazed by sheep). From 1988 or 1991, sheep were removed from one paddock/set. The other paddocks were grazed each summer, either moderately (3–4 sheep/ha) or heavily (10 sheep/ha). Vegetation was surveyed in summer 2010, in sixteen 30-cm-diameter circular quadrats/paddock. All quadrats were at a similar elevation (±10 cm). This study included the paddocks used in [11].

    Study and other actions tested
  13. A paired, site comparison study in 2010 in a salt marsh near Buenos Aires, Argentina (Di Bella et al. 2014) found that excluding livestock (along with legal protection) affected the plant community composition and species richness, with the effect depending on elevation, but did not significantly affect plant diversity. At high and medium (but not low) elevation, an exclusion area contained a significantly different plant community to a grazed area (data reported as a graphical analysis). This included greater relative cover of dominant denseflower cordgrass Spartina densiflora at the highest elevation (exclusion: 34%; grazed: 20%) and less relative cover of sea asparagus Sarcocornia perennis at the moderate elevation (exclusion: 9%; grazed: 29%). The exclusion area contained more plant species at the low and medium elevations (exclusion: 5–12; grazed: 3–9 species/transect) but fewer plant species at high elevation (exclusion: 28; grazed: 16 species/transect). At all elevations, plant diversity was statistically similar in exclusion and grazed areas (data reported as a diversity index). Methods: In spring 2010, vegetation was surveyed at six sites: three in a protected area from which cattle had been excluded for 30 years, and three in an adjacent grazed area (0.6 cattle/ha). Historically, all sites had been grazed at “very low” intensity. Note that this study evaluates the combined effect of excluding livestock and general legal protection. The sites were at high, medium or low elevation (i.e. flooded by tides twice yearly, twice monthly or twice daily). At each site, plant species and their cover were recorded along three 10-m-long transects.

    Study and other actions tested
  14. A site comparison study in 1988–2006 of coastal brackish and salt marshes in northern Germany (Rupprecht et al. 2015) reported that reducing grazing intensity (or stopping grazing entirely) affected vegetation development, but that the effect depended on multiple other factors. Grazing intensity was included as an important predictor in six of six statistical models of observed vegetation development (i.e. transitions between vegetation types over defined time periods). However, the effect of grazing intensity depended on other environmental conditions such as initial vegetation type, elevation and latitude (proxy for salinity and flooding frequency). In two models, for example, grazing intensity affected vegetation development at low but not high elevations. It is not possible to separate out results for reducing grazing intensity vs stopping grazing entirely. For example, moderately grazed plots showed similar responses to intensely grazed plots in some cases, but similar responses to ungrazed plots in others. Methods: In 1988, 1996, 2001 and 2006, plant community types were mapped across approximately 7,000 ha of brackish and salt marsh. Over time, grazing intensity was reduced in some areas, from intense (>10 sheep/ha) to moderate (≤3 sheep/ha) or zero. Where grazing was stopped, drainage systems were also abandoned but this had little effect on water levels. Statistical analyses were used to determine the influence of different factors, including grazing intensity, on changes in plant community types between the survey years.

    Study and other actions tested
  15. A replicated, paired, controlled study in 2002–2011 in a historically grazed brackish/salt marsh in the Netherlands (Chang et al. 2016) reported that the effects of excluding livestock on plant community composition, after nine years, depended on the elevation/wetness of plots. In higher, well-drained areas near to tidal creeks, exclusion plots developed a different plant community (dominated by sea couch grass Elytrigia atherica) to grazed plots (variable marsh communities). In lower, wetter areas further from creeks, a range of marsh plant communities developed in both exclusion and grazed plots with no clear distinction between the treatments. All data were reported as graphical analyses. The statistical significance of differences was not assessed. Methods: In spring 2002, twelve 10 x 25 m plots in a historically grazed coastal marsh were fenced to exclude livestock. The rest of the marsh remained grazed by cattle and horses during the summer. Regular tidal influx had been restored to the marsh over the previous five years. In summer 2011, cover of every plant species was estimated in seventy-two 4 x 4 m quadrats: three inside and three outside each exclosure.

    Study and other actions tested
Please cite as:

Taylor N.G., Grillas P., Smith R.K. & Sutherland W.J. (2021) Marsh and Swamp Conservation: Global Evidence for the Effects of Interventions to Conserve Marsh and Swamp Vegetation. Conservation Evidence Series Synopses. University of Cambridge, Cambridge, UK.

Where has this evidence come from?

List of journals searched by synopsis

All the journals searched for all synopses

Marsh and Swamp Conservation

This Action forms part of the Action Synopsis:

Marsh and Swamp Conservation
Marsh and Swamp Conservation

Marsh and Swamp Conservation - Published 2021

Marsh and Swamp Synopsis

What Works 2021 cover

What Works in Conservation

What Works in Conservation provides expert assessments of the effectiveness of actions, based on summarised evidence, in synopses. Subjects covered so far include amphibians, birds, mammals, forests, peatland and control of freshwater invasive species. More are in progress.

More about What Works in Conservation

Download free PDF or purchase
The Conservation Evidence Journal

The Conservation Evidence Journal

An online, free to publish in, open-access journal publishing results from research and projects that test the effectiveness of conservation actions.

Read the latest volume: Volume 21

Go to the CE Journal

Discover more on our blog

Our blog contains the latest news and updates from the Conservation Evidence team, the Conservation Evidence Journal, and our global partners in evidence-based conservation.


Who uses Conservation Evidence?

Meet some of the evidence champions

Endangered Landscape ProgrammeRed List Champion - Arc Kent Wildlife Trust The Rufford Foundation Save the Frogs - Ghana Mauritian Wildlife Supporting Conservation Leaders
Sustainability Dashboard National Biodiversity Network Frog Life The international journey of Conservation - Oryx Cool Farm Alliance UNEP AWFA Bat Conservation InternationalPeople trust for endangered species Vincet Wildlife Trust