Actively manage water level: brackish/salt marshes
Overall effectiveness category Likely to be beneficial
Number of studies: 10
Background information and definitions
This action involves active, repeated management of the amount of water in wetlands and when it is present, to mimic the natural hydrology of marshes. This may prevent excessively high or low water levels (e.g. during storm surges or droughts), or maintain wet/dry cycles that are a natural feature of many marshes (e.g. Zacharias & Zamperas 2010).
This action will usually involve some kind of water control structure: a valve, gate, sluice or pump. Water level management might aim to: mirror historical water level fluctuations or stability; manage salinity levels; increase sediment inputs; stimulate growth of desirable plant species; and/or create new wetland plant communities. Studies of “moist soil management”, “structural marsh management” and “environmental flows” along river courses all fall within the scope of this action. Caution: When managing water levels in a focal site, the effect on water levels in neighbouring sites should be considered.
Although water levels may be managed to restore or enhance habitats for waterfowl, information on the value of vegetation for waterfowl (e.g. seed production; productivity measured as CO2 exchange rates) is not summarized in this synopsis. Also, this synopsis does not include information on riparian areas that are not clearly marshes or swamps (e.g. riparian forests that require only a brief flood pulse for germination; Taylor et al. 2006).
Related actions: Raise water level to restore degraded marshes or restore/create marshes from other land uses; Lower water level to restore degraded marshes or restore/create marshes from other land uses; Facilitate tidal exchange to restore degraded marshes or restore/create marshes from other land uses; Manage water level to control problematic plants; Actively manage water level to complement planting.
Junk W.J., Piedade M.T.F., Schöngart J., Cohn-Haft M., Adeney J.M. Wittman F. (2011) A classification of major naturally-occurring Amazonian lowland wetlands. Wetlands, 31, 623–640.
Taylor J.P., Smith L.M. & Haukos D.A. (2006) Evaluation of woody plant restoration in the Middle Rio Grande: ten years after. Wetlands, 26, 1151–1160.
Zacharias I. & Zamparas M. (2010) Mediterranean temporary ponds. A disappearing ecosystem. Biodiversity and Conservation, 19, 3827–3834.
Supporting evidence from individual studies
A replicated, before-and-after, site comparison study in 1981–1984 of 12 adjacent brackish marshes in Manitoba, Canada (Welling et al. 1988) reported that during the first 2–3 months of drawdown following prolonged deep flooding, seedlings of dominant plants germinated. For four of five species, most seedlings grew around the elevation where adult plants had been dominant before flooding (see original paper for data). The exception was common reed Phragmites australis. Adult plants persisted at the higher elevations where common reed dominated, and these adult plants likely inhibited seedling growth. Most (81%) quadrats in the 10 experimental marshes contained seedlings of >1 species, whereas most (64%) quadrats in two nearby mature marshes contained adult plants of only one species. Methods: The water level in 10 slightly brackish (2–3 ppt) diked marshes on the shores of Lake Manitoba was actively managed: deep flooding for two years (water level raised 1 m above normal, killing most emergent vegetation) followed by drawdown in spring 1983 or 1984 (water level dropped to 20 cm below normal). This mimicked historical water level fluctuations in Lake Manitoba. Over the first summer of drawdown, seedlings were counted monthly in twenty 1-m2 quadrats/marsh. Pre-intervention vegetation was mapped from aerial photographs taken in August 1980. Plant species were also recorded in two adjacent, unmanipulated marshes in August 1983 (thirty 1-m2 quadrats/marsh). This study was based on the same experimental set-up as (2).Study and other actions tested
A replicated study in 1981–1984 of 10 adjacent brackish marshes in Manitoba, Canada (Welling et al. 1988) reported that seedlings germinated during two summers of drawdown following prolonged deep flooding, with most seedlings of dominant perennials germinating in the first summer. The study reported seedling numbers for seven herbaceous emergent species. For four perennial, grass-like species that dominated the marshes before intervention, 120–49,000 seedlings/100 m2 germinated in the first summer of drawdown (vs 160–3,400 seedlings/100 m2 in the second). For three annual forbs, 2,300–31,000 seedlings/100 m2 germinated in the first summer of drawdown (vs 85,000–200,000 seedlings/100 m2 in the second). Methods: The water level in 10 slightly brackish (2–3 ppt) diked marshes on the shores of Lake Manitoba was actively managed: deep flooding for two years (water level raised 1 m above normal, killing most emergent vegetation) followed by drawdown in spring 1983 or 1984 (water level dropped to 20 cm below normal). This mimicked historical water level fluctuations in Lake Manitoba. Seedlings were counted monthly in summer 1983 and 1984 in up to twenty 1-m2 quadrats/marsh. Quadrats were placed in the zone around the historical shoreline where emergent vegetation had been killed during flooding. This study was based on the same experimental set-up as (1).Study and other actions tested
A replicated, paired, site comparison study in 1989 in two brackish marshes in Louisiana, USA (Flynn et al. 1999) found that actively managing water levels within impoundments had mixed effects on the density and biomass of dominant saltmeadow cordgrass Spartina patens in each marsh. In Fina LaTerre marsh, saltmeadow cordgrass was always significantly less abundant in an impounded area than an area open to natural tidal exchange. This was true for density (impounded: 81; open: 99 stems/m2) and above-ground biomass (impounded: 897; open: 1,357 g/m2). In Rockefeller marsh, saltmeadow cordgrass abundance increased over the growing season. From August to November, it had a similar density in impounded and open marshes (impounded: 94–136 stems/m2; open: 107–117 stems/m2) and greater biomass in impounded marshes (impounded: 1,960–2,750 g/m2; open: 420–1,200 g/m2). The study suggests that the different responses in each marsh could be related to the design of the tidal control structures, distance of each marsh from the coast and soil chemistry. Methods: In 1989, and in each of two brackish marshes, vegetation was surveyed in an impounded area where water levels were managed (drawn down in spring/summer every 1–4 years, then reflooded in autumn/winter) and a nearby unmanaged area open to tidal exchange. Some parts of one marsh were also burned. Throughout the year, vegetation was cut from 0.1-m2 plots (7–19 plots/area/sampling date). Then, live cordgrass plants were counted, dried and weighed.Study and other actions tested
A replicated, site comparison study in 1996–1998 of 14 coastal brackish and salt marshes in Louisiana, USA (Gabrey et al. 1999) reported that active management of water levels within impoundments had mixed effects on winter vegetation cover, structure and species richness, depending on the year and whether marshes had been recently burned. In two of two years, impounded marshes had statistically similar overall vegetation cover (62–72%) to open marshes (56–78%). In one of two years, vegetation in impounded marshes created less visual obstruction than vegetation in open marshes (data reported as an index combining height and horizontal cover; other year no significant difference). Compared to open marshes, impounded marshes had similar or lower saltgrass Distichlis spicata cover (impounded: 0–2%; open: <1–11%), similar or higher cover of standing dead vegetation (impounded: 5–76%; open: 3–75%), and similar or higher plant species richness (impounded: 6–8 species/marsh; open: 4–6 species/marsh). Impounded marshes typically had higher saltmeadow cordgrass Spartina patens cover (three of four comparisons, for which impounded: 1–23%; open: <1–19%). Statistical significance of these cover results was not assessed. Methods: In January–February 1996 and 1997, vegetation was surveyed at 80 points in each of 14 brackish or saline marshes. Eight marshes had been impounded since the late 1950s, meaning water levels could be controlled (e.g. maintained relatively high in winter). Water levels were not controlled in the other six marshes (i.e. open to natural tidal influence). In each marsh, 40 points had been burned earlier in winter 1995/1996 and 40 had not. This study included the marshes studied in (6).Study and other actions tested
A replicated, paired, controlled, before-and-after study in 1989–1994 of 18 brackish marshes in southern France (Mesléard et al. 1999) reported that artificially flooded fields developed different plant communities with different species richness to fields with unmanaged flooding, but that the precise effects of artificial flooding depended on the flooding and/or grazing regime. Unless specified, statistical significance was not assessed. Over five years, the overall plant community composition changed in artificially flooded fields, becoming less like that of fields with unmanaged flooding (data reported as graphical analyses). Responses of individual plant species, and therefore the precise community that developed, depended on when fields were flooded and whether they were grazed. For example, final cover of common reed Phragmites australis was significantly greater in ungrazed, artificially flooded fields (12–16%) than in grazed, artificially flooded fields (<1%) or fields with unmanaged flooding (0%). In ungrazed fields, plant species richness was similar after five years of artificial flooding (5–6 species/0.25 m2) or unmanaged flooding (5 species/0.25 m2). In grazed fields, plant species richness was lower after five years of artificial flooding (4 species/0.25 m2) than unmanaged flooding (7 species/0.25 m2). Methods: The study used 18 adjacent former rice fields (1 ha; arranged in two sets of nine). From November 1989, six fields (three fields/set) received each flooding treatment: artificial winter flooding (10 cm depth November–April), artificial summer flooding (10 cm depth May–October) or year-round unmanaged flooding (inundated most of the winter most years). Half of the plots under each flooding treatment were also grazed. 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
A replicated, site comparison study in 1996–1998 of five coastal brackish marshes in Louisiana, USA (Gabrey et al. 2001) found that impounded marshes in which water levels were actively managed had similar summer plant species richness to open unmanaged marshes, and typically had similar cover of vegetation overall and dominant saltmeadow cordgrass Spartina patens. In three of three years, impounded marshes had statistically similar plant species richness (5.3–5.5 species/marsh) to open marshes (4.3–4.8 species/marsh). In two of three years, impounded marshes had statistically similar vegetation cover (total: 82–85%; cordgrass: 45–64%) to open marshes (total: 79–81%; cordgrass: 51–57%). The exception was in the first summer of the study, six months after prescribed burns, when impounded marshes had greater vegetation cover (total: 83%; cordgrass: 55%) than open marshes (total: 69%; cordgrass: 38%). Methods: In May–June 1996–1998, vegetation was surveyed at 80 points in each of five brackish marshes (salinity 5–10 ppt). Two marshes had been impounded since the late 1950s, meaning water levels could be controlled (e.g. maintained relatively high in winter). Water levels were not controlled in the other three marshes (i.e. open to natural tidal influence). In each marsh, 40 points had been burned earlier in winter 1995/1996 and 40 had not. This study used a subset of the marshes from (4).Study and other actions tested
A replicated, paired, before-and-after, site comparison study in 1991–1994 of four brackish marshes in Louisiana, USA (Johnson Randall & Foote 2005) found that active water level management typically had no significant effect on the height of the two dominant plant species, but had mixed effects on density. Before intervention, height and density of both saltmeadow cordgrass Spartina patens and American bulrush Schoenoplectus americanus were similar in all marshes. Over the three following years, managed and unmanaged marshes contained cordgrass stems of a similar height in eight of eight comparisons, and bulrush stems of a similar height in five of eight comparisons (data reported as height categories). In the other three comparisons, bulrush stems were shorter in managed marshes. Managed and unmanaged marshes contained a similar density of cordgrass in eight of eight comparisons (managed: 280–1,420 stems/m2; unmanaged: 350–1,520 stems/m2) but a lower density of bulrush in four of eight comparisons (for which managed: 20–60 stems/m2; unmanaged: 80–210 stems/m2). In three of the other four comparisons, bulrush density was similar in managed and unmanaged marshes. Methods: From 1992, water levels were controlled (drained in spring, rewetted in summer and flooded in autumn and winter) within two impounded marshes. In two adjacent marshes, the water level was not managed (no drawdown). Plant stems were counted and measured in each marsh five times before water management began (1991–1992) and eight times after (1992–1994). Some monitoring plots, both managed and unmanaged, were also fenced to exclude herbivores.Study and other actions tested
A before-and-after, site comparison study in 1996–2006 of four brackish/salt marshes in a delta in Texas, USA (Forbes et al. 2008) reported that adding treated wastewater to compensate for reduced natural freshwater inputs changed the overall plant community composition, and increased plant species richness. Unless specified, statistical significance was not assessed. Over the first four years after intervention, the overall plant community composition changed in the marsh directly affected by water additions, but was relatively stable in downstream marshes (data reported as a graphical analysis). In the marsh directly receiving water additions, sea marigold Borrichia frutescens cover significantly increased (before: <1–5%; after four years: 55%) whilst pickleweed Salicornia virginica cover significantly decreased (before: 83–88%; after: 34%). Meanwhile, in the downstream marshes, sea marigold cover did not significantly increase (before: 7–44%; after: 5–26%) whilst pickleweed cover did not significantly decrease (before: 28–54%; after: 28–56%). Plant species richness increased in the marsh directly receiving water additions (before: 1.6–2.1 species/1.25 m2; after four years: 3.4 species/1.25 m2) – but only to a similar level as the downstream marshes (before: 1.6–4.3; after four years: 3.1–4.4). Methods: Every day from October 1998, treated wastewater was discharged into one marsh in the Nueces Delta (via holding ponds), to compensate for reduced freshwater inputs and hypersalinity linked to upstream dams. Three downstream marshes, less affected by the wastewater inputs, provided comparisons. Plant species and their cover were surveyed before (June 1996–November 1997) and after (spring 1999–2002) water additions, along 11 transects/marsh/survey.Study and other actions tested
A replicated, paired, site comparison study in 2007–2009 of eight brackish/salt marshes in Texas, USA (Fitzsimmons et al. 2012) found that managed marshes (impounded and drawn down each spring/autumn, along with annual disking) and unmanaged marshes (subjected to neither of these interventions) had few plant species in common, but had similar overall plant species richness and diversity. Only 24–34% of plant species were found in both managed and unmanaged marshes (reported as a similarity index). However, both marsh types had statistically similar plant species richness (six of six comparisons; managed: 12–21 species/marsh; unmanaged: 8–18 species/marsh) and plant diversity (six of six comparisons; data reported as a diversity index). Methods: In autumn, winter and spring 2007/2008 and 2008/2009, vegetation was surveyed in four pairs of managed and unmanaged marshes (fifty-six 1-m2 quadrats/marsh, placed along transects). The managed marshes had been impounded for 6–9 years to control water levels and salinity (drawdown each spring-autumn) and the soil surface was disked every spring. The study does not distinguish between the effects of these interventions. All marshes were grazed each summer and burned every three years. The marshes were brackish in 2007/2008 (managed: <2 ppt; unmanaged: <10 ppt) but saline in 2008/2009 following a hurricane and storm surge (e.g. average salinity in managed marshes: 20 ppt).Study and other actions tested
A before-and-after study in 1994–2011 of a lakeshore brackish/saline marsh in Tunisia (Ouali et al. 2014) reported that after releasing water through upstream dams to restore winter flooding, the area of a bulrush-dominated community increased. In 1994–2002, the marsh was drier and more saline than normal. The main waterway feeding the marsh had been channelized to flow through the marsh (rather than into it) and dams had been built upstream. Vegetation dominated by bulrush Bolboschoenus glaucus covered 0–14 ha of the marsh, whilst salt marsh communities (dominated by glasswort Sarcocornia fruticosa and sea barley Hordeum marinum) covered 324–327 ha. In 2005, after three years of freshwater releases (along with heavy rains) that raised the water level of the lake, restored winter flooding and reduced salinity, bulrush-dominated vegetation expanded to cover 241 ha. Salt marsh vegetation covered 468 ha. In 2008–2011, continued freshwater releases (along with construction of dams and perpendicular embankments along the canal within the marsh, to hold back water; see Action: Raise water level to restore degraded marshes or swamps) maintained bulrush-dominated vegetation coverage at 192–298 ha. Salt-marsh vegetation covered 296–399 ha. Methods: Between 2005 and 2011, vegetation in Joumine Marsh was mapped using field surveys and satellite images. The study compared these maps to previously published maps. Large-scale release of freshwater into the lake began in autumn 2002.Study and other actions tested