Restore/create freshwater marshes or swamps (specific action unclear)

How is the evidence assessed?
  • Effectiveness
    not assessed
  • Certainty
    not assessed
  • Harms
    not assessed

Source countries

Key messages

  • Twenty-five studies evaluated the effects, on vegetation, of restoring/creating freshwater marshes or swamps using unclear or incompletely described actions. Twenty-three studies were in the USA. Two were in Canada. Two of the studies used the same set of wetlands.

VEGETATION COMMUNITY

  • Community types (1 study): One replicated, site comparison study in the USA reported that created wetlands had greater coverage of herbaceous vegetation after 7–8 years than natural wetlands, but lower coverage of forest and shrubby vegetation.
  • Community composition (17 studies): Four replicated, site comparison studies in the USA found that the overall plant community composition in created freshwater wetlands differed from the community in natural wetlands, after 1–21 years. Two replicated, site comparison studies in the USA and Canada reported mixed effects of freshwater marsh restoration/creation on overall algal or plant community composition, depending on the habitat and use of mining waste during creation. Of four replicated, site comparison studies in the USA and Canada, three reported lower quality vegetation in restored/created wetlands than in natural wetlands, but one reported similar vegetation quality in created and natural wetlands. Two replicated, site comparison studies in the USA found that created marshes developed a plant community characteristic of similar wetness to natural marshes within 4–21 years – but in one study, this was only true for created marshes >10 years old. Seven replicated studies in the USA simply quantified the composition, quality or wetness of the plant community up to 22 years after wetland restoration/creation.
  • Overall richness/diversity (17 studies): Eleven replicated studies, in the USA and Canada, compared overall plant richness/diversity in created/restored and natural/unmanaged freshwater wetlands. Five of the studies found that created/restored wetlands typically had similar plant taxonomic richness to natural/unmanaged wetlands. Three of the studies reported lower species richness in created than natural wetlands after 1–18 years. Two of the studies reported higher species richness in created than natural wetlands after 1–21 years. The final study reported mixed effects of marsh creation on plant species richness, depending on the vegetation zone and use of mining waste during creation. Two of the studies reported identical results for plant diversity as for richness (similar or greater in created vs natural wetlands) but one study found that the effect of management on plant diversity depended on the timing of drawdown. Six replicated studies in the USA simply quantified overall plant species richness and/or diversity over 1–16 years after wetland restoration/creation.
  • Native richness/diversity (3 studies): Of two replicated, site comparison studies of freshwater wetlands in the USA, one found that restored/created wetlands contained more native plant species than natural wetlands after 1–11 years. The other found that restored wetlands contained fewer native plant species than natural wetlands after 2–8 years. One replicated study of swamp restoration sites in the USA simply quantified native plant richness over 1–8 years after intervention.

VEGETATION ABUNDANCE

  • Overall abundance (7 studies): Six replicated studies, all in the USA, compared overall vegetation abundance in created/restored and natural wetlands. Four of the studies found that created/restored freshwater wetlands contained less vegetation (cover or biomass) than natural wetlands after 1–18 years. Two of the studies found that created and natural fresh/brackish/saline wetlands contained a similar amount of vegetation (overall cover and density; wetland plant cover) after >1 year. One of these studies reported that restored wetlands had lower vegetation cover than natural marshes – but this reflected management goals. One replicated study in the USA simply quantified total vegetation cover and biomass 3–10 years after marsh creation.
  • Herb abundance (2 studies): One replicated, site comparison study in the USA reported that created wetlands had greater overall cover of herb species, after 7–8 years, than natural wetlands. One replicated study in the USA simply quantified herb biomass in wetland restoration sites after 7–22 years.
  • Tree/shrub abundance (1 study): One replicated study in the USA simply quantified the density of woody vegetation in wetland restoration sites after 7–22 years.
  • Algae/phytoplankton abundance (1 study): One replicated, site comparison study in the USA found that ≤15-year-old restored freshwater marshes contained a similar phytoplankton biomass to natural marshes.
  • Individual species abundance (9 studies): Nine studies quantified the effect of this action on the abundance of individual plant species. For example, one replicated, site comparison study in the USA found that created and natural freshwater marshes supported a similar abundance of pickerelweed Pontederia cordata after 1–11 years.

VEGETATION STRUCTURE

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, site comparison study in 1993 of 20 freshwater marshes in Florida, USA (Streever et al. 1996) found that stands dominated by pickerelweed Pontederia cordata had similar vegetation cover and density in created and natural marshes. After 1–11 years, pickerelweed-dominated stands contained a statistically similar abundance of vegetation in created marshes and natural marshes. This was true for pickerelweed cover (created: 54%; natural: 55%) and density (created: 157 stems/m2; natural: 164 stems/m2). Ignoring one unusual natural site, the same was also true for non-pickerelweed cover (created: 20%; natural: 39%) and density (created: 31 stems/m2; natural: 35 stems/m2). Twelve plant species were recorded in pickerelweed stands across all the created marshes (vs 15 species in natural marshes). Methods: In summer 1993, vegetation was surveyed in pickerelweed-dominated stands within ten marshes created after mining, and ten natural marshes (fifteen 0.5-m2 quadrats/marsh). All marshes were 1–80 ha in area. Created marshes had been excavated between 1982 and 1992. Some were probably amended with wetland soil and/or planted with wetland plants, although details of wetland creation are not clear. The study did not attempt to measure the overall vegetation composition of the marshes, only that of pickerelweed stands.

    Study and other actions tested
  2. A replicated study in 1994 of five freshwater wetland restoration/creation sites in Ohio, USA (Wilson & Mitsch 1996) reported that four of the sites contained the desired emergent vegetation within four years, and that wetland-characteristic species made up at least 45% of the vegetation cover in each site. Restoration/creation was intended to replace areas of emergent wetland vegetation. After approximately 1–4 years, four of five sites did contain emergent wetland vegetation (other site: submerged vegetation surrounded by upland). Across all sites, wetland areas contained 4–52 plant species (excluding “unidentified grasses”). Wetland species made up 65­–100% of vegetation cover. Wetland-characteristic species made up 45­–100% of vegetation cover. For data on the absolute cover of individual plant species, see original paper. Methods: In 1994, vegetation was surveyed in 0.25-m2 quadrats (number not clear) in five wetland restoration/creation sites. The sites were approximately 1–4 years old. Interventions were not clearly reported by included excavation, rewetting or other management of water inputs and outputs, planting herbs, planting trees/shrubs, and invasive species removal. Surveys included emergent, floating and submerged vegetation in wetland habitats only, and where water was <1 m deep.

    Study and other actions tested
  3. A replicated, site comparison study in 1993 of 96 freshwater wetlands in Oregon, USA (Magee et al. 1999) found that restored/created wetlands contained a different overall plant community to natural wetlands, with more plant species, after 1–11 years. The overall plant community composition differed significantly between restored/created and natural wetlands (data reported as graphical analyses). Although 220 of 365 recorded plant taxa occurred in both restored/created and natural wetlands, 86 occurred only in the former and 59 only in the latter. Amongst restored/created wetlands, the overall plant community composition changed over time: wetlands >3 years old contained a significantly different community to wetlands ≤3 years old. Finally, restored/created wetlands contained more plant species (overall: 41; native: 19; non-native: 19 species/wetland) than natural wetlands (overall: 30; native: 13; non-native: 15 species/wetland). For data on the frequency of some individual plant species, see original paper. Methods: In summer 1993, plant species were recorded in 96 wetlands (approximately fifty-seven 1-m2 quadrats/wetland). Fifty-one wetlands had been restored or created 1–11 years previously. The study does not report details of restoration/creation methods. Forty-five wetlands were naturally occurring. All wetlands were ≤2 ha in size and were dominated by wet meadows, emergent marshes, floating vegetation and/or open water. Restored/created wetlands were wetter than natural wetlands, with 52% of their area flooded during the growing season (vs 21% in natural wetlands).

    Study and other actions tested
  4. A replicated, paired, site comparison study in 1997 of 69 restored/created wetlands in Massachusetts, USA (Brown & Veneman 2001) found that they contained different plant communities to remnant natural wetlands, with fewer plant species and lower cover. After 3–12 years, restored/created sites contained a plant community characteristic of wetland conditions (data reported as a wetland indicator index). However, the overall plant community composition typically differed between restored/created wetlands and remnant natural wetlands (similarity <38% in 66 of 69 cases; statistical significance not assessed). The restored/created wetlands also had fewer plant species (restored/created: 9; natural: 11), fewer wetland plant species (restored/created: 8; natural: 10), and lower vegetation cover (all plants and wetland plants; data not reported). Methods: In summer 1997, vegetation was surveyed in 69 pairs of wetlands. Each pair included one wetland restoration/creation project and one remnant natural wetland. One representative 15 x 30 m plot was surveyed in each wetland. The restoration/creation projects were intended to compensate for damage to wetlands from development projects. The study included both marshes and swamps (number of each not clearly reported).

    Study and other actions tested
  5. A replicated, site comparison study in 1995 of 26 freshwater marshes in Pennsylvania, USA (Campbell et al. 2002) found that created marshes developed a wetland-characteristic plant community after 10 years, but had lower plant species richness and vegetation cover than reference (disturbed and pristine) marshes for up to 18 years. The overall plant community in >10-year-old created sites was characteristic of true wetlands, just like the community in reference wetlands. However, the overall plant community in <10-year-old created sites was characteristic of significantly drier conditions (data reported as a wetland indicator index). Regardless of age, created wetlands had a lower average plant species richness (3–6 species/79 m2) and vegetation cover (65–75%) than reference wetlands (10 species/79 m2; 95% cover). Methods: In June–August 1995, vegetation was surveyed across 26 freshwater marshes. Twelve marshes (0.2–5.3 ha) had been created in uplands approximately 2–18 years previously. The study does not report details of wetland creation methods. Fourteen reference marshes (0.1–2.1 ha), ranging from “disturbed” to “pristine”, were used for comparison. Plant species were recorded in 5-m radius circular plots, spaced 20 m apart across the whole of each wetland. Vegetation cover was recorded in a 0.25-m2 quadrat within each plot.

    Study and other actions tested
  6. A replicated, site comparison study in 1995 of 53 prairie pothole wetlands across four complexes in North Dakota, USA (Mushet et al. 2002) found that restored wetlands contained lower quality vegetation and fewer native plant species than natural wetlands after 2–8 years, but a similar total number of plant species to natural wetlands after 5–8 years. In all three restored complexes, the native wetland vegetation was of lower quality (i.e. less characteristic of undisturbed local habitats) than in a natural complex (data reported as conservatism scores and a floristic quality index). However, vegetation quality increased with time since restoration. In all three restored complexes, wetlands contained fewer regional native plant species (27–34 species/wetland) than wetlands in the natural complex (44 species/wetland). However, in the two oldest restored complexes, wetlands contained a similar number of plant species in total (50–51 species/wetland) to wetlands in the natural complex (56 species/wetland). For data on the frequency of individual plant species, see original paper. Methods: In summer 1995, plant species were recorded in 53 depressional wetlands. There were 11–14 wetlands in each of three complexes restored two, five and eight years previously. Restoration methods were not clearly reported, but involved reseeding grasslands around the wetlands. No native wetland species were planted. Sixteen wetlands were within a natural, relatively undisturbed wetland complex. Restored wetlands spanned a similar range of sizes and flooding regimes to the natural wetlands.

    Study and other actions tested
  7. A replicated, site comparison study in 2001–2002 of 15 freshwater wetlands in West Virginia, USA (Balcombe et al. 2005) reported differences in the area and cover of vegetation between created and natural marshes after 4–21 years. Results summarized for this study are not based on assessments of statistical significance. Created wetlands were 0% forest by area (vs natural wetlands: 5%), only 5% shrubs (vs natural: 41%), 54% herbaceous (vs natural: 44%) and 41% open water (vs natural: only 9%). Plant species with different cover in created and natural wetlands included reed canary grass Phalaris arundinacea (created: 7%; natural: 0%), common rush Juncus effusus var. effusus (created: 5%; natural: <1%), tussock sedge Carex stricta (created: <1%; natural: 4%) and broadleaf cattail Typha latifolia (created: 1%; natural: 7%). Methods: In summer 2001 and 2002, vegetation was surveyed in fifteen wetlands: eleven created/restored 4–21 years previously (details not reported) and four natural (undisturbed). Coverage of vegetation types was estimated across each wetland, and cover of individual species was estimated in at least five 1-m2 quadrats/vegetation type/wetland. This study used the same sites as (8).

    Study and other actions tested
  8. A replicated, site comparison study in 2001–2002 of 15 freshwater wetlands in West Virginia, USA (Balcombe et al. 2005) found that created wetlands contained different, richer and more diverse vegetation to natural wetlands, but with a similar proportion of wetland species and total herb cover. After 4–21 years, both created and natural sites contained wetland-characteristic plant communities (data reported as a wetland indicator index). However, the plant species composition significantly differed between created and natural wetlands (data reported as a graphical analysis). Created wetlands supported a greater abundance of non-native and early-colonizing species (e.g. reed canary grass Phalaris arundinacea; see Study 7 and original paper for cover data). Created wetlands had greater plant species richness on average (13 species/500 m2) than natural wetlands (8 species/500 m2). The same was true for diversity (data reported as a diversity index). In contrast, created and natural wetlands contained a statistically similar percentage of wetland-characteristic species (created: 79%; natural: 90% of all plant species) and total cover of herb species (created: 39%; natural: 54%). Methods: In summer 2001 and 2002, vegetation was surveyed in fifteen wetlands: eleven created/restored 4–21 years previously (details not reported) and four natural (undisturbed). Plant species were recorded in at least one 500-m2 plot/vegetation type/wetland. Cover of these species was estimated in five 1-m2 quadrats/plot. This study used the same sites as (7).

    Study and other actions tested
  9. A replicated study in the early 2000s of 45 created, restored and enhanced freshwater wetlands in the USA (Spieles 2005) reported that they typically contained wetland vegetation after 1–7 years. In 43 of the 45 studied wetlands, the overall plant community was more characteristic of wetlands than uplands. The other two wetlands contained vegetation that was marginally more characteristic of uplands than wetlands (data reported as a wetland indicator index). On average, the wetlands contained 11 plant species/10 m2 (range: 0.2–56 species/10 m2), of which 16% were not native (range: 0–53%). The study also reported effects of intervention type (creation, restoration or enhancement), wetland setting (depressional, riverine or lakeshore) and wetland age on these metrics, and reported data on the frequency of some individual plant species (see original paper). Methods: Data on vascular plant species and cover were collected from monitoring reports for 45 wetlands (within 36 project areas across 21 states). The wetlands were 1–562 ha in area and 1–7 years old. Of the 45 wetlands, 17 had been created (from scratch), 19 restored (from degraded wetlands) and nine enhanced (increasing value of existing wetlands) as part of mitigation projects. Thirty-two wetlands contained areas of marsh and 19 contained areas of shrubby/forested swamp. The study does not consistently separate results for marsh and swamp vegetation.

    Study and other actions tested
  10. A replicated, site comparison study in 2000 of 45 freshwater marshes in Michigan, USA (Zheng & Stevenson 2006) found that restored marshes had lower cover of emergent vegetation and duckweed than natural marshes, a different community of algae growing on plants, and greater algal richness and diversity. The restored marshes were ≤15 years old. On average, restored marshes had lower vegetation cover than natural marshes. This was true for emergent canopy cover (restored: approx. 20%; natural: approx. 40%) and duckweed cover (restored: 36%; natural: 59%). The overall community composition of algae growing on plants differed significantly between restored and natural marshes, but the plankton and sediment algal communities did not (data reported as graphical analyses). Species richness of all three algal groups was statistically similar in restored marshes (9–30 taxa/marsh) and natural marshes (8–26 taxa/marsh). Phytoplankton biomass was also statistically similar in restored and natural marshes (data not reported). Methods: In July 2000, vegetation was surveyed in 25 restored marshes of varying age (restored ≤15 years ago) and 20 naturally occurring marshes. The study does not report details of restoration methods. All marshes were permanent or semi-permanent, surrounded by farmland, <2 ha in area and <2 m deep. Surveys included emergent vegetation, floating vegetation, phytoplankton (two samples/marsh from 10 cm below water surface), algae growing on plant stems (4–5 stems/marsh) and algae growing in the sediment (8 samples/marsh from the top 1 cm).

    Study and other actions tested
  11. A replicated, site comparison study in 2004–2005 of 18 freshwater marshes in Oklahoma, USA (Hartzell et al. 2007) found that created marshes contained a greater proportion of perennial plant species than natural marshes, but that other vegetation metrics were similar in created and natural marshes. After >20 years, 94% of the plant species in created marshes were perennial – significantly greater than the 78% of perennial species in natural marshes. All other vegetation metrics were statistically similar in each type of marsh. These included taxonomic richness (created: 6.9; natural: 6.9 taxa/marsh), taxonomic diversity (data reported as a diversity index), proportion of wetland-characteristic species (created: 77%; natural: 67%), vegetation quality (data reported as an index based on how characteristic each plant species is of undisturbed local habitats) and percent cover of the dominant species (created: 55%; natural: 49%). Methods: In summer 2004 and 2005, herbaceous plant species and their cover were recorded in six created wetlands (>20 years old; identified based on the presence of a dam or levee, and soil characteristics) and 12 natural wetlands. All wetlands were <1.1 ha. The study does not report further details of wetland creation methods. The created wetlands had deeper, less turbid water than the natural wetlands and were typically flooded for longer. Vegetation was surveyed in four 0.5-m2 quadrats/broad vegetation type/wetland/year.

    Study and other actions tested
  12. A replicated study in 1992–2004 of 11 wetland restoration sites in Wisconsin, USA (Nedland et al. 2007) reported that they developed wetland-characteristic herbaceous vegetation, which increased in quality over time. Each site was surveyed 1–4 years and 13–16 years after restoration. Nine sites contained mostly emergent vegetation, but two contained mostly submerged vegetation. Between the two surveys, the proportion of wetland-characteristic plant species increased, from 54% to 81% of all species present. The plant species also became more characteristic of undisturbed local habitats, on average (data reported as a conservatism score and floristic quality index; both increased but only the former significantly). However, there were declines in total plant species richness (from 108 to 84 species across all wetlands, and from 29 to 20 species/15 m2) and plant diversity (reported as a diversity index). For data on the presence/absence of individual plant species, see original paper. Methods: In August 1992 and 2004, plant species and their cover were recorded in 1-m2 quadrats in each of 11 wetland restoration sites (each <1 ha; 15–30 quadrats/site/survey). A complete plant species list was also made for each wetland. Restoration interventions had been carried out between 1988 and 1991 (details not reported). The analysis was based on change within each wetland over time.

    Study and other actions tested
  13. A replicated, site comparison study in 2001 of 19 freshwater marshes in Ohio, USA (Fennessy et al. 2008) found that created marshes contained less vegetation biomass than natural marshes. After 1–9 years, created marshes contained less above-ground vegetation biomass on average (209 g/m2) than nearby natural marshes (347 g/m2). Methods: In July and August 2001, live above-ground vegetation was collected 19 depressional marshes (eight 0.1-m2 quadrats/marsh), then dried and weighed. Ten marshes had been created 1–9 years previously. The other nine marshes were natural. The study does not report the specific interventions used for marsh creation, but does note that excavation was involved in all created sites.

    Study and other actions tested
  14. A replicated, site comparison study in 2007 of 35 freshwater, brackish and saline marshes in Hawaii, USA (Bantilan-Smith et al. 2009) found that restored/created marshes had statistically similar overall plant species richness, cover of wetland plant species and cover of exotic plant species to natural marshes, but that restored marshes had lower overall vegetation cover. Data were not reported for most outcomes. Overall vegetation cover was only 59% in restored marshes: significantly lower than the 74% cover in created wetlands and 76% cover in natural wetlands. The study suggests this may reflect management of restored wetlands for waterbirds, which require open water and mudflats. Methods: In March/April 2007, plant species and their cover were recorded in 35 coastal lowland marshes: 11 restored, 7 created and 17 natural. Six 1-m2 quadrats were surveyed in each marsh, across flooded, saturated and upland areas. Eight marshes were freshwater, whilst the other 27 were influenced by salt (brackish–hypersaline). The study does not report details of restoration/creation methods (including dates), or separate the effects of restoration/creation in marshes of different salinity.

    Study and other actions tested
  15. A replicated study in 1991–2006 of 15 restored marshes and nine restored swamps in Illinois, USA (Matthews & Endress 2010) found that the overall plant community composition changed over the four years following restoration, and reported that the wetlands overall experienced a net gain in plant species richness. The plant community in younger restored marshes was characterized by annual herbs like Pennsylvania smartweed Persicaria pensylvanica and barnyard grasses Echinocloa spp., whereas older restored marshes were characterized by clonal perennial herbs like broadleaf cattail Typha latifolia, reed canarygrass Phalaris arundinacea and rice cutgrass Leersia oryzoides (data reported as a graphical analysis). Across all 24 restored marshes and swamps, plant species richness increased over time. On average, each wetland lost 20% of the plant species present at the start of each year but gained 27%. Methods: Between 1991 and 2002, twenty-four degraded wetlands were restored by (a) removing drainage structures, building berms and/or excavating, then (b) planting herbs, trees, shrubs, or nothing. The study does not describe which intervention(s) were carried out in each wetland. Vascular plant species (and, in marshes, their cover) were recorded in the first four summers following restoration.

    Study and other actions tested
  16. A replicated study in 2010 of 53 wetland restoration/creation sites in the southeast USA (De Steven & Gramling 2012) reported that they developed stands of native wetland vegetation after 2–11 years. The sites had developed into herbaceous, shrubby or forested habitats (or a mixture of these). The overall plant community in restored/created sites was characteristic of undisturbed local wetlands (data reported as a wetland indicator index and conservatism score). Of the 21–31 dominant plant species in each site, at least 86–91% were wetland species (capable of growing in wetlands) whilst 59–69% were wetland-characteristic species (that always or usually grow in wetlands rather than uplands). Between 93 and 96% of species were native. These averages mask the fact that in 6–8 of the 53 sites, the overall plant community was characteristic of drier conditions (based on the wetland indicator index) and/or <40% of plant species were wetland-characteristic. Methods: Dominant plant species (>20% cover in any layer of vegetation) were surveyed in 53 wetland restoration/creation sites in July–August 2010. Most or all sites were probably fresh rather than saline (not explicitly reported). Restoration or creation was completed 2–11 years previously using interventions such as ditch blocking, excavation, relandscaping and tree planting. The study does not report details of restoration/creation methods, or separate results for marsh and swamp vegetation.

    Study and other actions tested
  17. A replicated study in 1989–2005 of 15 created wetlands in Virginia, USA (DeBerry & Perry 2012) reported that they had developed vegetation cover within 1–15 years of creation. On average, there were 29 plant species/site in the herbaceous layer and 6 plant species/site in the shrub/sapling layer. The herbaceous layer was dominated by perennial species (71% of species; 69% importance). The shrub/sapling layer was predominantly volunteer species (52% of species; 67% importance). None of these metrics significantly differed between sites of different ages. The same was true for plant diversity. Although there was typically no significant difference in overall plant community composition between sites of different ages, there was a general pattern of older communities (11–15 years old) returning to a composition more like younger communities (1–5 years old) after an intermediate (3–10 years old) shift in composition (data reported as graphical analyses and similarity indices). The dominant species also varied between sites of different ages (see original paper for data). Methods: In late summer 2004 or 2005, vegetation was surveyed in fifteen 1–17 ha depressional floodplain wetlands. All wetlands had been created 1–15 years previously with the intention that they would develop into swamps (precise details not reported or not available, but involving planting trees/shrubs and sowing grass to stabilize the soil). The herbaceous layer (herbs and trees/shrubs <1 m tall) was surveyed in fifteen 1-m2 quadrats/wetland. The shrub/sapling layer (plants >1 m tall) was surveyed in one 5-m-radius plot/wetland.

    Study and other actions tested
  18. A replicated study in 2009 of four created freshwater marshes in Virginia, USA (Dee et al. 2012) reported that they contained wetland vegetation after 3–10 years. In all four marshes, the overall plant community was characteristic of wetland conditions (data reported as a wetland indicator index). There were 4–5 plant species/m2 and 19–27 plant species/marsh. Of these, 70–84% were wetland-characteristic and 20–26% had been sown. Overall vegetation cover was 99–106%, including 24–65% sown species and 1–26% non-native. Above-ground vegetation biomass was 770–1,830 g/m2. Most tested metrics did not significantly differ between wetlands of different ages: overall cover, sown cover, species richness and diversity, and vegetation quality. However, older wetlands did have a plant community characteristic of slightly drier conditions, and the oldest wetland supported the lowest vegetation biomass (see original paper for data). Methods: The four wetlands were created between 1999 and 2006 (details not reported, except that some herb seeds were sown). Vegetation was surveyed in August or September 2009 (16–32 quadrats/marsh). Plant species and their cover were recorded across the whole of each 1-m2 quadrat. All standing vegetation was cut from 0.25 m2 subquadrats, then dried and weighed.

    Study and other actions tested
  19. A replicated, site comparison study in 2008–2009 of 54 ephemeral freshwater marshes in Mississippi, USA (Fleming et al. 2012) found that intervention to create waterfowl habitat had no significant effect on plant species richness or the relative abundance of woody plants, and only affected plant diversity and the relative abundance of grasses if it involved early drawdown. At the end of the growing season, managed and unmanaged marshes contained a similar number of plant taxa (managed: 17; unmanaged: 16 taxa/marsh) and a similar proportion of woody plants (3% of all plants). Managed marshes with an early drawdown of the water table had greater diversity of plant species than unmanaged marshes (data reported as a diversity index), a greater variety of plant growth forms (1.8 different forms/sample point; vs unmanaged: 1.4) and a greater proportion of grasses (47% of all plants; vs unmanaged: 26%). These metrics did not significantly differ between managed marshes with a later drawdown of the water table and unmanaged marshes (see original paper for data). Methods: In October 2008 and 2009, vegetation was surveyed at 50–64 points in each of 54 marshes across 18 private lands. The vegetation of 39–42 marshes had been actively managed for waterfowl, with interventions such as annual soil disturbance, herbicide or mechanical control of undesirable (low forage value) plants and summer drawdown. Drawdown was early (before 15 June) in approximately half of the managed marshes, and late (≥3 weeks later) in the other managed marshes. The final 12–15 marshes received almost no intervention (e.g. “limited or no control of undesirable plants”).

    Study and other actions tested
  20. A replicated, paired, site comparison study in 2007–2011 of 133 fresh and brackish wetlands in Alberta, Canada (Wilson et al. 2013) found that restored/created wetlands typically contained lower quality wet meadow vegetation than natural wetlands. All data were reported as a floristic quality index, with higher quality vegetation containing species more characteristic of undisturbed wet meadows in the study area, and a greater proportion of native species. In six of eight comparisons, vegetation in restored/created wetlands was of lower quality than in natural wetlands. This was true for wetlands restored on historically mined land (including some still affected by pollution) and for “improved” stormwater ponds (with attempts to make them more like natural wetlands using interventions such as reprofiling, adding wetland soil and planting). In the final two comparisons, vegetation in wetlands restored on farmland contained was of similar quality to vegetation in natural wetlands. Methods: Between 2007 and 2011, plant species were recorded in the wet meadow zone of 133 wetlands (six 1-m2 quadrats/wetland). There were 47 restored or created wetlands (≥ 3 years old) and 86 naturally occurring wetlands (some surrounded mostly by forest, some mostly by agriculture) across two distinct regions. The study does not report full details of restoration/creation methods, or separate results for fresh and brackish wetlands.

    Study and other actions tested
  21. A replicated, site comparison study in 2008–2009 of four floodplain wetlands in Texas, USA (Wall & Stevens 2014) found that created wetlands contained a different plant community to a natural wetland, with fewer wetland plant species. After 7–8 years, the plant community composition in the created wetlands was only 14–35% similar to the natural wetland. In three of four created wetlands, the proportion of wetland species was significantly lower (20–87%) than in the natural wetland (96%). Wetland plant species richness was lower in created than natural wetlands in seven of eight comparisons (for which created: 0.3–0.5 species/m2; natural: 2.0–3.3 species/m2; other comparison no difference). Total plant species showed mixed results depending on the wetland and year: similar in created and natural wetlands in four of eight comparisons (created: 2.0–4.5 species/m2; natural: 2.0–4.3 species/m2) lower in created wetlands in three comparisons (created: 0.5–2.8 species/m2; natural: 2.5–4.2 species/m2) and higher in created wetlands in one comparison (created: 3.8 species/m2; natural: 2.5 species/m2). For data on the presence/absence of individual plant species, see original paper. Methods: In summer/autumn 2008 and 2009, plant species were recorded in four created wetlands and one nearby natural wetland (sixty 1-m2 quadrats across all sites). The study does not clearly report the interventions used for wetland creation, but does note that culverts were installed in 2001 to allow water flow between the wetlands, and that no vegetation was introduced. However, the created wetlands were ephemerally flooded whilst the natural wetland was permanently flooded.

    Study and other actions tested
  22. A replicated study in 2012 of thirty wetland restoration sites in Illinois, USA (Jessop et al. 2015) reported that they developed vegetation within 7–22 years, but this was characteristic of disturbed habitats. The wetland restoration sites contained 13–755 g/m2 dry above-ground herb biomass (average: 265) and 0–1 woody stems at least 1 m tall/m2 (average: 0.3). They had 0–95% cover of non-native plants (average: 36%) and a conservatism score of 2.1–3.3 out of 10 (average: 2.7), indicating that the plant species were characteristic of fairly disturbed habitats in the study region. Methods: In summer 2012, vegetation was surveyed in 30 wetland sites, restored 7–22 years previously. Restoration involved excavating, removing drainage structures, building embankments to retain water and, in some sites, planting and/or sowing wetland plants. The precise interventions carried out in each site are not clearly reported. Of the 30 sites, 13 were marshes, 10 were swamps and seven included both marshy and swampy areas. The study does not separate results for marsh and swamp vegetation.

    Study and other actions tested
  23. A replicated, site comparison study in 2008–2012 of marsh vegetation in 51 freshwater wetlands in Alberta, Canada (Roy et al. 2016) reported mixed effectiveness of wetland creation on the overall plant community composition and species richness, depending on whether mining waste was incorporated into the wetlands and on the vegetation zone. For wetlands created without incorporating mining waste, the emergent zone supported a statistically similar plant community to natural wetlands (80% similarity), but had lower species richness (created: 5.9; natural: 8.6 species/6 m2). In contrast, the wet meadow zone supported a significantly different plant community to natural wetlands (70% similarity), but had statistically similar species richness (created: 9.7; natural: 9.9 species/6 m2). For wetlands where mining waste was used during creation, the plant community composition in both zones was significantly different from natural wetlands (50% similarity). Species richness in the emergent zone (3.9 species/6 m2) was lower than in the natural wetlands. Species richness in the wet meadow zone (9.7 species/6 m2) was statistically similar to the natural wetlands. The study identified physical environmental differences between wetlands that may have influenced the vegetation. Methods: In August 2008–2012, vegetation was surveyed in 51 wetlands near Fort McMurray: 19 created on natural sediment and with fresh surface water; 16 created with mining waste (built on waste sediments, or filled with waste water); 16 natural. All created wetlands were ≥7 years old (18 were >15 years old). The study does not report details of wetland creation methods. Vegetation (emergent, floating and submerged) was surveyed in six 1-m2 quadrats/vegetation zone/wetland.

    Study and other actions tested
  24. A replicated, paired, site comparison study in 1996–2012 involving 15 wetland restoration sites in Illinois, USA (van den Bosch & Matthews 2017) found that they had higher or lower quality vegetation than nearby natural wetlands, depending on the metric. The 15 restoration sites had lower quality vegetation than 15 nearby natural wetlands when measured as a conservatism score (how characteristic the plant species are of undisturbed local habitats), but higher quality vegetation when measured a floristic quality index (combining conservatism score with species richness). The proportion of native, non-weedy, perennial species was statistically similar in the restoration sites (49–55% of species) and the natural wetlands (60% of species). Methods: The study involved 15 sites restored between 1992 and 2004 (methods not reported) and a natural wetland site near each restored site. The study included both marshes and swamps (number of each not clearly reported). Vegetation was surveyed twice in restoration sites (once in 1996–2009, once in 2012) and once in natural wetlands (2012).

    Study and other actions tested
  25. A replicated study in 2009 of 13 swamp restoration sites in the USA (Wood et al. 2017) reported that they had developed vegetation cover, including native species, after 1–8 years. The restoration sites contained 8–36 native plant species, which comprised 28–100% of the overall vegetation cover. The number and abundance of woody plant species was not reported. Additional analyses did not separate results for wetland and upland sites. Methods: In summer 2009, plant species and foliage cover were surveyed in 13 sites being restored as freshwater forested wetlands (one 405-m2 plot/site). Restoration began 1–8 years previously. The study does not report the restoration methods in detail, but they included controlling invasive plants that were initially dominant and covering the ground surface (with cardboard, wood chippings or landscape fabric), then regular management by watering, introducing native plants and protecting them from herbivores.

    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 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, terrestrial 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 18

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 Programme Red List Champion - Arc Kent Wildlife Trust The Rufford Foundation Save the Frogs - Ghana Bern wood Supporting Conservation Leaders National Biodiversity Network Sustainability Dashboard Frog Life The international journey of Conservation - Oryx British trust for ornithology Cool Farm Alliance UNEP AWFA Butterfly Conservation People trust for endangered species Vincet Wildlife Trust