Netherlands: Restoration of the Scheldt Estuary’s Sieperda Tidal Marsh

Project Stage:
Implementation

Start Date:
1990-11-10

End Date:
1990-11-10

Primary Causes of Degradation

Degradation Description

The estuary has been in an almost continuous state of change during the 20th century. Because of land reclamation, reinforcement of dikes, and the deepening of the shipping channels, about 2,500 hectares of mudflats and tidal marshes have been lost since 1900. Substantial parts of the remaining marshes have reached the final stage of succession. Sheltered areas with the potential to develop into intertidal areas and marshes have become very scarce.

Reference Ecosystem Description

The estuary contains important habitats for shellfish (cockles) and juvenile flatfish (plaice and flounder), and it provides important habitat for many species of birds. Migratory birds use the area for foraging, and a number of coastal breeding birds nest along the estuary.

Project Goals

To understand the forces driving ecological change in the Sieperda marsh since the dike breach in 1990

Monitoring

The project does not have a monitoring plan.

Description of Project Activities:
In 1966, a sand dam was constructed on the Sieperda marsh parallel to the existing sea dike to accommodate gas and water pipelines. The area between the dam and the sea dike was closed off from tidal influence by the construction of a relatively low (summer) dike, turning it into a polder, 85% of which was leveled and used for cattle grazing and agriculture. Only the easternmost part of the polder, closest to the estuary, kept its natural geomorphologic structure. A long ditch was dug along the sand dam for drainage of rainwater and occasional floodwater. Several smaller ditches were linked onto the main drainage system. Because the summer dike was relatively low, the area frequently flooded in winter. In 1976 and 1985 the dike gave way during sever storms but was subsequently repaired. After another dike breach in 1990, it was decided not to repair the dike and allow the area to be returned to tidal influence once again. The accidental breach turned the Sieperda polder into a developing tidal marsh. This project did not have a specific restoration goal but offered an opportunity to learn how the return of tidal influence would have its effect on a stretch of reclaimed land. During the past 10 years, four independent studies were conducted, dealing with changes in hydrology and geomorphology, vegetation structure, benthic fauna, and bird populations. The propagation of the tidal wave (velocity and amplitude) was measured during complete neap-spring tidal cycles using pressure sensors (PTX 630, Druck, The Netherlands) at three locations. A pressure sensor was placed on the bed in the center of the creek. Tidal surveys were performed in 1992, 1993 (twice), 1995, and 1996 (Sánchez Leal et al. 1998). The developments in creek and marsh morphology were studied by conducting field surveys and by interpreting aerial photographs. Sedimentation and erosion rates in creeks and on the marsh surface were monitored along six transects, five transects perpendicular and one transect parallel to the direction of tidal flow. Altimetry measurements were performed along the transects by leveling on an annual basis from 1992 onward (Komman & van Doorn 1997; Sánchez Leal et al. 1998). All bed levels were measured relative to Dutch Ordnance Datum (N.A.P.). From the changes in bed levels over time sedimentation rates on the marsh were calculated. Creek and tidal flat development was also determined by analyzing aerial photographs taken in 1990 and 1995. The monitoring of developments in vegetation succession commenced in 1993. At the start of the study, 29 permanent quadrats (PQs, 2 x 2m) were marked along seven transects, which were evenly distributed across the marsh. The composition of the vegetation within each PQ was surveyed annually in September. For each plant species, the percentage cover was estimated using a percentage gradient. Marsh areas reflecting less than 25% of maximum reflectance were considered uncovered or sparsely covered. The sampling of benthic macrofauna started in 1995 and was performed annually in September, the end of the growth season when maximum biomass was expected. Ten sampling sites were selected that were evenly distributed over those parts of the marsh not covered with vegetation. Six sites were situated on mudflats and four in shallow supralittoral pools. Through the years, some sampling sites became overgrown with vegetation, and these sites were excluded from the study. Infauna was sampled by taking ten 30-cm deep sediment cores at each site with a round metal corer (diameter 4.5cm). The cores were sieved over 1-mm round mesh and stored in 4% neutralized formalin. Hyperfauna, in this case the fauna living near the bottom in the shallow water of the pools, was sampled with a hyperbenthic sampling net (width 25cm, mesh 1.25mm). Each sample was the result of a 20-m haul. The residue was preserved in 4% neutralized formalin. Each sample was analyzed in the laboratory to species level, where possible using a binocular microscope; and biomass (ash free dry weight) was determined by drying (80ºC, 48 hr) and ashing (570°C, 2 hr) each sample (Stikvoort 2000). The significance of the developing marsh for breeding birds was studied by monitoring and mapping the exact nest sites in 1991, 1994, 1997, and 1999 using standard procedures (Hustings et al. 1985). During the breeding season, the area was visited at least six times, and all observations of breeding were recorded. All water birds in the marsh were monitored on a monthly basis as part of an existing long-term monitoring program. The number of water birds (grebes, herons, geese, ducks, and waders) were counted or estimated around high tide, using binoculars and telescopes, when the birds concentrated at high-tide roosts.

Ecological Outcomes Achieved

Eliminate existing threats to the ecosystem:
In the eastern part of the Sieperda marsh, geomorphologic changes were very pronounced. Creeks became wider and deeper over time, and levees developed. The sedimentation rate on creek levees was high and constant, while the sedimentation rate in the floodplain areas of the marsh changed with time. During the first years after the construction of the creek, the sedimentation rate was higher (3 cm/yr) than average for Scheldt estuary marshes, which is approximately 1.5 cm/yr (Krijger 1993). The new creek, constructed in 1993, allowed larger tidal volumes into the developing marsh, causing increased creek erosion and enhanced sediment availability. After this period of strong sedimentation, the sedimentation rate of the floodplains approached values characteristic for the estuary. The period 1997-1999 was characterized by the strong growth of Phragmites australis (common reed), a dense and high vegetation very capable of sediment trapping. This is thought to be the main cause for the increased sedimentation rate observed during this period. In the western part of the marsh, hardly any geomorphologic changes occurred. The sedimentation rates were less than 0.5 cm/yr, and creeks did not develop. The large difference in development between the eastern and western marsh is caused by the weak tidal influence in the western marsh, which was unable to alter the geomorphology. After the reintroduction of tidal influence, the former crop fields, as well as those parts of the former pastures with poor drainage, turned into bare mudflats or areas very sparsely covered with vegetation. Vegetation succession on the mudflats was initiated with the establishment of the salt marsh species Aster tripolium and Puccinellia maritima, followed after several years by Scirpus maritimus and Atriplex prostrata. The former pastures were initially characterized by common grasses such as Agrostis stolonifera, Elymus pycnanthus, and E. repens. In the low-lying and moist areas, A. stolonifera was replaced almost completely by the salt marsh species P. maritima and A. tripolium after the return of tidal influence. In the higher and drier parts of the former pastures, A. stolonifera and E. pycnanthus could survive and were accompanied by marsh plants such as P. maritima, A. tripolium, and Juncus gerardi. In closely cropped areas in the westernmost distant part of the marsh with insufficient drainage, pioneer species such as A. tripolium, Salicornia spp., Spergularia marina, and Glaux maritima could flourish. From 1995 onward, the colonization of the mudflats progressed rapidly, probably as a result of improved drainage after the construction of the new creek in 1993. In recent years, Phragmites australis became established in the marsh and expanded rapidly. In 1995, the first year of monitoring and 5 years after the dike breach, estuarine benthic fauna was well established in the newly developed marsh. A total of 19 taxa was observed in the marsh over the period 1995-1999. The observed insects and fish (goby) inhabited the supralittoral pools. Species diversity between years varied from 14 to 17 taxa. Species densities fluctuated between 4,000 and 40,000 individuals/m², with an average of some 10,000 individuals/m². Total biomass fluctuated between sampling sites and years and varied between 1 and 27g ash-free dry weight/m². Relative species abundance also fluctuated over time, although the two most abundant species (Corophium volutator and Nereis diversicolor), often accompanied by several oligochaete spp., constituted more than 80% of total species density at most sampling sites. In fact, C. volutator and N. diversicolor alone accounted for more than 80 to 90% of total biomass at most sites. The return of tidal influence in Sieperda marsh appears to have affected breeding birds, because breeding habitats changed as a result of vegetation development and succession. Not surprisingly, a number of bird species that breed on salt marshes, such as Anser anser, Tringa tetanus, and Luscina svecica, have taken advantage of the new situation. Acrocephalus scirpaceus, a species breeding on marshes or reed beds, was also able to benefit. On the other hand, species whose feeding habitat is restricted to mudflats (such as Dunlin) declined in numbers as vegetation succession progressed. The number of Oystercatchers, a species that breeds in sparsely vegetated areas, also declined with increasing vegetation.

Factors limiting recovery of the ecosystem:
After the dike breach in 1990, the tidal flow into the polder was restricted at first due to the existing drainage network. In 1993, a new wider creek was dug in the center of the area, enhancing tidal influence into the developing marsh. Although tidal influence into the marsh improved, the distal part of the marsh experienced tidal influence only during spring tides and showed a spring to neap tidal cycle instead of a diurnal cycle as observed in the proximal part of the marsh. A typical tidal curve shows a slightly asymmetric sinusoidal wave, with a shorter flood than ebb duration. The tidal range in the estuary near Sieperda marsh, including the easternmost part of the marsh, varied from 4 to 6m. After the construction of the new creek and bridge in 1993, the average tidal range behind the bridge increased from 0.75 to 1.75m. Three years later, the average tidal range had increased to 2.3m. Tidal influence further into the marsh is hampered by the width of the bridge acting as a bottleneck and the very slow growth of the creek in the central part of the marsh due to the presence of a solid surface layer of clay. As a result, there is no creek penetration into the western part of the marsh, and tidal volumes in this area are small.

Socio-Economic & Community Outcomes Achieved

Key Lessons Learned

The reintroduction of tidal influence, and the resulting geomorphologic changes, directed the ecological changes in the former polder. It can be concluded that the restoration of the Sieperda marsh has been successful, at least in the eastern (seaward) half of the marsh. Developments in the western half, both geomorphologically and ecologically, were less pronounced due to a hampered tidal influence. Widening the bridge and enlarging the creek through the clay layer would improve tidal influence in this area.

Because Sieperda marsh once was part of the adjacent Saeftinge marsh, the marsh surface level (more than mean high water level) at the time of the dike breach was typical of an old salt marsh. Therefore, vegetation succession commenced rapidly and within 10 years reached the stage of the emergence of common reed (P. australis). A more gradual and natural rate of succession could be achieved by leveling elevated parts of the marsh, which would rejuvenate the process of marsh development and stimulate the settlement of pioneer vegetation.

Long-Term Management

Salt marsh restoration through depoldering, while shown to be successful in this and other projects from northwestern Europe, is a contentious issue that demands a careful balance of competing interests: flood control and coastal protection, maintenance of port facilities, and ecosystem vitality and longevity. Because the public is still accustomed to the idea of reclaiming land from the sea, and because tidal marshes are not as highly valued as agricultural fields, public support for projects such as these has been limited to date. Thus, the long-term success of this project, and others like it, depends upon the ability of restoration planners and practitioners to garner public support and demonstrate the utility of restored coastal systems as an important means of improving coastal resilience and defense.

Primary Contact

Organizational Contact