With the decisions about the Development outline 2010 and the updated Sigmaplan, the Dutch and Flemish governments are committed to leap forward with the ecological rehabilitation of the Scheldeestuary. An important challenge is the creation of tidal wetlands; the transformation of woods or agricultural land into tidal mudflats and marshes. In order to assess the feasibility and to identify possible problems any similar small scale projects already in place should be studied in detail. Analysis of their evolution can improve our apprehension of the larger scale future plans. This report deals with the monitoring results of three small scale tidal wetland restorations in the Zeeschelde. The restoration measures as well as the monitoring strategies are evaluated. A fourth project was finished only recently and is discussed only briefly.
In 2003 Ketenisse schor, in the brackish part of the estuary, was separated from the inland by the construction of a new dike and the area itself was lowered back under mean high water level. An open brackish tidal area of 60ha (35.5ha restored) was created. A similar measure was taken in 2004 on the Paardenschor, in order to restore 12ha of brackish tidal wetland. However, this site resembled more a breached site because some existing tidal marsh remained between the river and the restored area. Also in 2004, a long and narrow fresh tidal zone with terraces (1.6ha) was created near Paddebeek, in a part of the estuary where tidal wetlands are scarce. In Heusden a fresh tidal wetland of 10ha was created through realignment in 2006. The old dike was lowered to mean high water level and two breaches to mean low water level were excavated where the old drainage sluices used to be.
At each of the project sites changes in geomorphology, sediment characteristics, sediment quality, vegetation, macrozoobenthos, avifauna and fish were studied in the first years after restoration. These developments were compared as much as possible to the situation on nearby tidal wetlands.
Sedimentation/erosion was evaluated on three different spatial scales: locally with sedimentation/erosion plots, along trajectories perpendicular to the river with RTK-GPS and full cover with LIDAR and aerial photography. Areas under MHW with a gentle slope were subject to net sedimentation. The sedimentation rate was directly proportional to shelter (width of the area on open tidal areas and distance from the breach in breached areas). Areas with a steep slope were subject to erosion. A critical overall slope of 2.5% was calculated. Above this slope erosion was more likely to take place. Depressions acted as sediment traps and filled relatively quickly. Zones above mean high water level showed very little geomorphologic changes in this short study period. Local patterns were related to resistant layers on the surface, such as clay banks. Sedimentation/erosion was the overall result of slope, intertidal elevation, width, shelter and soil properties. In some combinations no net
sedimentation or erosion was observed, but both processes alternated in time, probably according to the composition and dynamics of the inundating water. In this study only characteristics of the restored site were considered. To validate these conclusions in a wider context it would be necessary to take local estuarine characteristics such as local sediment budget and wave energy into account. Dendritic and sinuous creek systems only developed in zones with net sedimentation. Creek development was found to be directly proportional to the mean width of the area perpendicular to the river. More higher order creeks developed in wider areas. An exponential and inverse relationship between slope and creek development was found. In open realignments separate parallel creek systems developed whilst in a breached situation fewer but bigger systems with higher order creeks developed. In future monitoring more attention should be paid to differences between breached systems and open areas.
In the restored sites with net sedimentation the soil developed from sand to fine sands or silt. On locations where erosion and sedimentation alternated, a high variability in median grain size was observed but no correlation between median grain size and the occurrence of sedimentation or erosion could be established. However, a negative correlation between erosion and chlorophyll a content was found. On most locations organic matter percentage increased but only in the extremely sheltered locations it became comparable to that of existing mudflats. The study period was too short to evaluate the time needed to establish fully functional tidal wetland soils. The expectation is that the organic matter content will build up at a rate that is inversely proportional to local dynamics. A negative correlation between median grain size and organic matter content and chlorophyll a was found as well as a positive relation between organic matter and chlorophyll a in the upper sediment layer. Several studies report an inverse correlation between median grain size and tidal elevation. This gradient could not be established in this study. This is probably related to geomorphologic diversity as well as
the short observation period since restoration. Moreover, local differences in dynamics also play a role. This eventual gradient should be examined on a more local scale and in more uniform conditions. An altitude gradient in chlorophyll a concentrations was found, related to exposure time and potential photosynthetic activity. Organic matter and chlorophyll a content were found to be higher in spring and summer.
Immediately after the restoration of the Paardenschor and Ketenisse schor the calculated scores for total pollution on the restored areas and the original mudflats were comparable. Even sampling locations where very little sedimentation or net erosion took place also showed high levels of total pollution from the very first sampling campaign. However, the sediment of the original mudflat at Ketenisse schor showed higher concentrations of cadmium and mercury. Occasionally significant differences in sediment quality were observed between two nearby sampling locations, despite corrections made for silt and organic matter content when calculating the quality scores. The general expectation or fear for tidal wetland restoration in the Zeeschelde is that sediment quality on the restored sites would deteriorate through deposition of contaminated sediment. However, at most sampling stations the score for total pollution did not change over time. Only individual
contaminant concentrations changed considerably at some sampling locations. In Heusden recently deposited sediment showed significantly higher concentrations of organic contaminants whilst the original soil was more contaminated with heavy metals.
The restored areas were partly colonized by macro-algae and higher plants during the first growing season. A well defined distinction was found between plant communities on restored sites situated in the brackish part and the freshwater part of the river Scheldt. In the brackish part 11 vegetation types were distinguished, whereas in the fresh water part there were 6 types. This is mainly due to the larger surfaces and the marked differences in topography and inundation frequency at the restored sites in the brackish part. In general, locations situated high in the tidal frame were colonized most rapidly. However, significant differences were observed between the sampling stations and not all changes could be linked to tidal elevation. The stability and succession of the vegetation and the colonization rate were strongly related to geomorphological processes, which were in turn directed by the condition of the area just after restoration. On marshes with a steep slope erosion and regressive succession were observed. Areas with continued sedimentation and a more gentle slope showed progressive succession.
Benthic invertebrates rapidly colonized the restored areas. Only one month after restoration several taxa were found. The number of taxa quickly increased. On the Paardenschor the number of taxa remained highest on the original mudflat throughout the study period. On Ketenisse schor on the other hand, the number of taxa was higher on the restored area than on the original mudflat after October 2004. This was probably more related to the large differences in habitat characteristics between the locations on the restoration site and on the original mudflat. Mainly mobile taxa appeared shortly after restoration in the brackish part, particularly Nereis diversicolor
and Corophium volutator
. Less mobile species remained rare throughout the study period. It is expected that the populations of those species will become larger in the future. In total about 20 benthic invertebrate taxa were found at the Paardenschor and Ketenisse schor. The number of benthic taxa at Paddebeek was small and species composition differed from that at the brackish restoration sites. Only Paranais litoralis
was found on all sites. Both abundance and biomass of benthic invertebrates were low at the start of the monitoring but quickly increased throughout the study period. Until spring 2005 abundances and biomasses were lower on the restored area of the Paardenschor than on the original mudflat. Afterwards they compared to each other. The low abundances observed on the original mudflat at Ketenisse schor were probably related to the dynamics of the habitat type involved. C. volutator
and N. diversicolor
made the largest contributions to total biomass. Densities of Oligochaeta remained relatively low on the restored sites. This is possibly related to the absence of a planctonic larval stage in this species group. At Paddebeek low abundances of Oligochaeta were found immediately after restoration but they quickly increased significantly. As soon as marsh vegetation colonised the higher zones, snails such as Alderia modesta
and Assiminea grayana
followed, as well as herbivorous Tulipid and Limoniid insect larvae.
Birds occupied the restored areas almost immediately after restoration. In the first years the distinct wintering seasonal pattern of the Zeeschelde was absent on the restored sites.
A total of 19 species of water birds were recorded at the Paardenschor. Common Shelduck, Mallard, Curlew and Oystercatcher were almost always observed. The restored site is used for foraging as well as roosting. Benthic invertebrates are abundant and the inundation time of the newly created mudflat is limited. Only few suitable breeding sites were available throughout the study period. The importance of the Paardenschor as a breeding ground is expected to increase as parts of it develop into tidal marsh.
At the Ketenisse schor a total of 46 water bird species were recorded. After three seasons a wintering pattern started to develop. Ducks, especially Shelduck, were the most abundant species but also waders and geese were numerous. Waders and Shelduck mainly foraged on the wider muddy areas in the central zone of the restored area, where benthic invertebrates were most abundant. The numbers of roosting birds were also highest on these areas. The highest numbers of geese, mostly Greylag geese, were observed in winter. They primarily foraged on the restored areas with Scirpus maritimus. At Ketenisse schor the number of breeding waders, particularly Avocet and Little ringed plover, decreased and the numbers of reed birds increased as higher vegetation established.
In the brackish restored areas fish utilised the restored sites for foraging at high tide. In the freshwater restored sites remaining pools also seem to function as spawning and nursing habitat for tolerant species such as Prussian carp and Stone moroko.
Evaluation of the restoration design:
Paardenschor: The initial tidal elevation and site slope were well chosen. Creek network systems seem to establish without the specific excavation of a creek onset. However, this might have enhanced the habitat differentiation within the site and its suitability as fish habitat. The old dike might have been excavated more, but this might have led to erosion on the transition to the Schor Ouden Doel. Overall there is net sedimentation, with local erosion in the developing creek network system. The sediment is colonised by benthic invertebrates, predated on by water birds and fish. Higher vegetation is developing slowly. The site is functional as a roost and foraging site rather than as a breeding site. This type of low dynamic mudflats, relatively high in the tidal frame adds valuable foraging time and space for water birds. However, its design could have been adapted to enhance its habitat functions for fish.
Ketenisse schor: The slopes at the extreme ends of the site are too steep and net erosion takes place. In the central part two aspects of the final design differed significantly from the original plan. Some areas, where the topsoil was not useful as construction material for dikes, were not excavated below mean high water level and remained almost supratidal. The old dike was not removed according to plan and as a result almost flat plateaus, with a steep slope towards the river were constructed instead of a gentle overall slope from the dike to the river. This had consequences for the habitat functions of the site. At T0 higher vegetation was already in place, supratidal as well as tidal marsh vegetation. Some of it died off, in other places it remained and served as source for typical fauna and flora. The plateaus now provide low dynamic habitat. They silted up and a relatively rich macro-benthic invertebrate community was built up, providing extra foraging and roosting time and space for birds. The steeper parts of the site showed net erosion and seemed less functional as habitat. If the central
slope would have been excavated according to plan sedimentation/erosion and habitat development would have been quite different. There would have been relatively less mudflat with a long exposure time. On the other hand, habitat diversity and gradual transitions might have been more elaborate.
Paddebeek: Through the inland shifting of the dike a small tidal area could develop in an area of the river Schelde where mudflats and marshes are scarce. Because of the construction of terraces with willow wicker, stone rubble was not necessary to protect the new dike. Unfortunately the greater part of the old dike remained in place, hindering proper drainage, creek formation and colonisation. To allow some drainage several stones should be removed. The terraces were constructed with life willow wicker. As a consequence, willow shrubs established very quickly, which accelerated vegetation succession. The site has limited habitat functions for birds. Nevertheless, it is valuable for the connectivity of the tidal wetlands in this part of the Zeeschelde.
Heusden LO: The new inland dike is not fortified with stone rubble, and the topography of the restored site was not altered. The old dike was not lowered to MLW level as planned but rather to MHW. Initially the site only inundated at spring tides and it was not drained at low tide. Later two breaches to MLW were added where the old sluices used to be, connecting to ditches. It then had every aspect of a breached site with a strongly accentuated spring tide/neap tide differentiation in the inundation regime. Nevertheless, some areas remained inundated at low tide and the southern part where the sand stock for the dike construction works was not completely removed remained supratidal. As a result of this design a site with a great variety of habitat types was created, with permanent pools, mudflats and all stages of a typical tidal marsh vegetation. The vegetation gradient from low marsh to supratidal was uninterrupted because of the absence of fortifications. Unfortunately the area was recently colonised by Floating marsh pennyworth, an invasive species. Chances are that this species will soon invade the complete tidal area.
Evaluation of the monitoring design:
The monitoring design was evaluated based on the results and analyses and it was compared to the strategies of three British restoration sites.
Generally in the Zeeschelde more effort was put in site success monitoring on the site itself and relatively too little attention was paid to impact verification monitoring and ‘reference measurement’ of identical variables on established tidal wetlands with similar characteristics. Because of the often ‘ad hoc’ situation monitoring schemes were more according to the possibilities on the spot. For the larger scale projects of the Development outline 2010 and the updated Sigmaplan monitoring should be part
of the project planning and budget. Reference and impact verification monitoring should start well before the restoration works. A yearly fixed point photographic survey would help to keep track of the changes over time, to visualise them and to facilitate the interpretation of the collected data. The extensive gamut of tools (LIDAR, aerial views, sedimentation/erosion plots, etc.) used to evaluate morphological changes meet the needs to monitor morphological changes on different scales. For future projects the results of these exercises should be used to select the most appropriate combination of tools, according to the situation. Due to the high costs and relatively long time needed to plan the financing of LIDARs and aerial views, availability of those tools cannot always be guaranteed. Alternatively laser altimetry by means of a Riegl-scanner can be used if the geometry of
the site is adequate. Profiles and surveys with RTK-GPS also seemed adequate to follow-up global changes in slope and tidal elevation. Local morphological changes were monitored with sedimentation/erosion plots or plates. Monthly measurements of the sedimentation/erosion plots and plates proved to be sufficient.
Sampling sediment characteristics such as median grain size and organic matter content simultaneous with sedimentation/erosion processes is recommended to analyse the relation between those variables and processes. Monitoring the chlorophyll a content is recommended to evaluate the colonisation with microphytobenthos as it plays an important role in sediment stability and as a food source for several species of benthic invertebrates and water birds. The set of measured sediment characteristics should ideally be expanded with soil density, water content, redox potential and sulphide content.
Contrary to the monitoring on the British restoration projects, not only heavy metals but also organic pollutants were measured. This additional information proved to be very useful. The set of pollutants should remain the same in future monitoring campaigns. Sampling the top 2cm horizon instead of the top 10cm will be considered. This may yield a better reflection of the quality of recently deposited sediment. The number of sampling points for sediment quality can be restricted to a few for each site. So far only few floristic data were collected. These data are generally used to determine marsh quality. Comparison of floristic data with data collected in permanent quadrates will reveal whether this needs elaboration.
For a better interpretation of macrobenthic data sampling should take place in the same month each year and for every site. Increasing the sampling frequency is not feasible due to the labour-intensive process of handling the samples.
Monitoring the water birds on a monthly base seems to be the maximum that is possible. The monitoring, however, should be on a more structural base and according to a fixed protocol.