Remove surface soil/sediment: freshwater marshes
Overall effectiveness category Awaiting assessment
Number of studies: 6
Background information and definitions
Surface soil/sediment – and any vegetation on it – could be removed to create a new bare surface for plants to colonize. This new surface may have fewer nutrients and pollutants, have no undesirable seed bank, and have a looser surface. Soil/sediment removal can also make a site wetter, by bringing the surface closer to the water table, increasing the frequency/duration of tidal flooding, or increasing the water depth in an already flooded site. This action may be particularly useful in naturally dynamic habitats that have been artificially stabilized, mimicking disturbances that would create bare soil/sediment.
Caution: Heavy machinery is usually needed for this action. Heavy vehicles can churn and compress wetland soils (Campbell et al. 2002). Stripping topsoil can have counter-intuitive effects, such as increasing ammonium concentrations because nitrifying bacteria, which break down ammonia, are removed with the soil (Dorland 2004). It may remove seeds of desirable species, and can be expensive.
Related actions: Raise water level to restore degraded marshes; Raise water level to restore/create marshes from other land uses; Reprofile/relandscape; Bury surface soil/sediment; Disturb soil/sediment surface without removing material; Transplant or replace wetland soil; Remove surface soil/sediment before planting.
Campbell D.A., Cole C.A. & Brooks R.P. (2002) A comparison of created and natural wetlands in Pennsylvania, USA. Wetlands Ecology and Management, 10, 41–49.
Dorland E. (2004) Ecological restoration of wet heaths and matgrass swards: bottlenecks and solutions. PhD Thesis, Utrecht University, The Netherlands.
Supporting evidence from individual studies
A replicated, site comparison study in 1993–1998 involving 12 dune slacks in the Netherlands (Grootjans et al. 2001) reported that slacks where topsoil was removed (along with stopping groundwater extraction and reintroducing grazers) developed plant communities with characteristic wetland species and more plant species than mature, unmanaged slacks. Statistical significance was not assessed. Restored slacks developed plant communities, the overall composition of which changed over time (data reported as a graphical analysis). After five years, restored slacks contained 76–108 plant species overall and 48–86 species/100 m2. This included species characteristic of dune slacks (5–11 species/100 m2) and nutrient-rich marshes (2–11 species/100 m2) alongside other wetland and upland species. In each slack, total vegetation cover was always <50% and only two individual species – creeping willow Salix repens and bushgrass Calamagrostis epigejos – ever had cover >1%. For comparison, during the second year of the study, mature slacks contained 12–39 plant species/m2 (data not reported for other outcomes). Methods: Dune slacks are low-lying areas amongst dunes. Eight degraded slacks (stabilized and covered with undesirable, mature vegetation) were restored. In summer 1993, vegetation and topsoil were removed (10–40 cm depth, across all or part of each slack). Earlier that year, groundwater extraction had been stopped. In 1995, grazers (a “small herd” of cattle and ponies) were reintroduced to seven slacks. The study does not distinguish between the effects of these interventions. Vegetation was surveyed in at least five of the restored slacks (spring or summer 1994–1998) and four mature slacks (spring 1994): species across the whole of each slack; species and cover in five comparable 100-m2 plots/slack.Study and other actions tested
A study in 1989–1994 in a freshwater marsh in Florida, USA (Dalrymple et al. 2003) reported that areas where topsoil was removed were colonized by vegetation, with species richness and the amount of tall/shrubby vegetation depending on the amount of topsoil removed. After approximately 54 months, an area where topsoil had been completely removed contained 32 plant species/100 m2, 243% total vegetation cover and 79% cover of wetland-characteristic plants. The formerly-dominant shrub Brazilian pepper Schinus terebinthifolius occurred in only 4% of survey plots. Less than 1% of total cover was plants >2 m tall. An area where topsoil had been partially removed contained 20 plant species/100 m2, 245% total vegetation cover and 81% cover of wetland-characteristic plants. Brazilian pepper occurred in 86% of survey plots. Approximately 10% of total cover was plants >2 m tall. Results were similar 30–42 months after soil removal, although there was some variation in the first 6–18 months (see original paper). Methods: In early 1989, topsoil and vegetation were removed from a marsh that had been farmed and then became overgrown with Brazilian pepper. Topsoil was completely removed (down to bedrock) from 18 ha and partially removed (“thin layer” remaining) from an adjacent 6 ha. Plant species and cover were recorded each August between 1989 and 1994, in fourteen or forty-nine 100-m2 plots/area/year.Study and other actions tested
A replicated, site comparison study in 2007–2008 in an overgrown floodplain wetland in central Japan (Ishii et al. 2011) found that some plots stripped of topsoil and vegetation were colonized by new marsh vegetation, and that these plots contained more plant species over two growing seasons than adjacent unstripped land. Unless specified, results summarized for this study are not based on assessments of statistical significance. Vegetation colonized shallow-stripped plots (flooded 22–57 days/year) but not deeper stripped plots (flooded 215 days/year). Over the first two growing seasons, the stripped plots contained more plant species (102) than adjacent unstripped land (66). Sixty-five species only occurred in the stripped plots. The stripped plots also contained more plant species characteristic of wetlands/wet disturbed floodplains (stripped: 37; unstripped: 8), but a statistically similar number of alien plant species (stripped: 9; unstripped: 3). After two growing seasons, the shallow-stripped plots contained 0.7–11.1 plant species/m2 (including 0.3–2.2 wetland-characteristic), 7–167 plants/m2 (including 0.8–11.3 wetland-characteristic) and <5–33% vegetation cover. Invasive goldenrod Solidago altissima was absent from all stripped plots. Methods: In spring 2007, topsoil and vegetation were removed from ten 70-m2 plots on a goldenrod-invaded floodplain (two plots for each of five stripping depths; 1.5–2.7 m of topsoil removed). Vascular plants were surveyed between spring and autumn 2007 and 2008. All species were recorded in the stripped plots, plus cover and density in nine 1-m2 quadrats/plot/survey. Species were also recorded along transects in unstripped land within 50 m of stripped plots.Study and other actions tested
A site comparison study in 2008 of five sedge meadows in Illinois and Wisconsin, USA (Lawrence & Zedler 2013) found that a meadow restored by removing excess sediment (and trees, then planting tussock sedge Carex stricta) – contained more but smaller sedge tussocks than nearby natural meadows after 11–14 years. In four of four comparisons, the restored meadow contained a greater density of sedge tussocks (8.4 tussocks/m2) than natural meadows (4.5–5.6 tussocks/m2). Sedge tussocks were also smaller in the restored meadow than in the natural meadows. This was true in four of four comparisons for height (restored: 5 cm; natural: 11–18 cm), perimeter (restored: 39 cm; natural: 51–82 cm) and volume (restored: 560 cm3; natural: 2,342–6,604 cm3). The basal area of tussocks in the restored meadow was only 0.07 m2/m2, compared to 0.12–0.23 m2/m2 in the natural meadows (statistical significance not assessed). Methods: In 2008, sedge tussocks were surveyed in one restored and four natural sedge meadows (15–30 quadrats/meadow, each 1 m2). The restored meadow was formerly a wooded floodplain. Trees and accumulated sediment were removed, then plugs of tussock sedge planted 30 cm apart, between 1994 and 1997. The study does not distinguish between the effects of these interventions on any non-planted sedges.Study and other actions tested
A replicated, site comparison study around 2010 of 48 ephemeral freshwater marshes in Nebraska, USA (O’Connell et al. 2013) reported that marshes undergoing restoration (agricultural topsoil removed and surrounding cropland abandoned) contained a different plant community to natural marshes (surrounded by permanent grassland) and degraded marshes (surrounded by cropland), with lower cover of wetland perennial plants and fewer wetland perennial species than the natural marshes. Results summarized for this study are not based on assessments of statistical significance. After 1–12 years, the overall plant community composition differed between restored, natural and degraded marshes (data reported as a graphical analysis). Perennial wetland species were underrepresented in restored marshes (43% cover; 10.1 species/marsh) compared to natural marshes (56% cover/group; 13.0 species/marsh). However, restored marshes had greater cover of these species than degraded marshes (35% cover; richness not reported). Annual wetland species were “slightly” overrepresented in restored marshes compared to natural marshes in terms of abundance (data reported as a graphical analysis). However, there was a similar number of these species in restored and natural marshes (8.2 vs 8.0 species/marsh). Methods: Around 2010, vegetation was surveyed in 48 ephemeral playa marshes (along two transects crossing each marsh, in both the cool and warm seasons). Sixteen of the marshes were undergoing restoration under the Wetland Reserve Program. This involved removing eroded agricultural topsoil from the marshes and abandoning the surrounding cropland. The study does not distinguish between the effects of these interventions. Of the remaining marshes, 16 were in natural catchments and 16 were in degraded, farmed catchments.Study and other actions tested
A replicated, site comparison study in 2010 of 39 prairie pothole wetlands in North Dakota, USA (Smith et al. 2016) found that restoration by excavating excess sediment (and sometimes planting wetland herbs) reduced cover of hybrid cattail Typha x glauca, but that other effects on vegetation depended on the vegetation zone. Across both the marsh and wet meadow zones, restored potholes had lower hybrid cattail cover (6%) unrestored potholes (19%). In the marsh zone, the overall plant community composition significantly differed between restored and unrestored potholes (data reported as a graphical analysis). Restored potholes also had less horizontal vegetation cover (data not reported). In the wet meadow zone, neither the plant community composition nor horizontal vegetation cover significantly differed between restored and unrestored potholes. Compared to natural potholes, the restored potholes had a significantly different plant community in both zones and lower horizontal cover in the marsh zone, but similar horizontal cover in the wet meadow zone and similar hybrid cattail cover (natural: 5%). Methods: In summer 2010, vegetation was surveyed in the marsh (seasonally flooded) and wet meadow (occasionally flooded) zones of 39 prairie potholes (10 quadrats/zone/pothole). Thirty potholes were surrounded by former cropland, converted to perennial vegetation cover. Excess cropland sediment had been removed from 19 of these potholes, 2–7 years previously. Prairie cordgrass Spartina pectinata had also been planted in the wet meadow zone of some excavated potholes (number not reported). The study does not distinguish between the effects of sediment removal and planting on any non-planted vegetation in these potholes. The remaining nine potholes were “natural”, i.e. surrounded by land that had never been cultivated.Study and other actions tested