Restoration of species-rich grasslands on reconstructed river dikes

Up until 30 years ago an extensive, flower-rich grassland vegetation containing many species rare in the Netherlands used to be common on Dutch river dikes. However, the deterioration of the flora on dikes was already being reported at the end of the 1960s. At that time too, ecologists warned that the planned reinforcement of the dikes along the Rhine, Waal, Lek and IJssel would adversely affect the flora. Their gloomy forecasts have proved to be correct. Between 1968 and 1992 as much as 89% of the locations with a dry floodplain grassland vegetation in the Netherlands disappeared.In 1992 the vegetation of over 90% of the river dikes consisted of species-poor grassland on which sheep graze, and rough vegetation mown for hay. Only about 7% of the surface area of the river dikes was covered by relatively species-rich grasslands belonging to the phyto-sociological syntaxa Arrhena-theretum elatioris and Lolio-Cynosuretum , both belonging to the Arrhenatherion elatioris . Only 1% was covered by the typical species-rich dry grassland Medicagini-Avenetum . The last remnants of these grasslands are in imminent danger of disappearing.The deterioration in the semi-natural vegetation has mainly been caused by the fact that the slopes of the dikes are increasingly being used agriculturally (fertilization, overgrazing, use of herbi-cides) but also because ecological features were insufficiently taken into account while reinforcing the dikes. In 1984 a research project was started to ascertain the optimum structure and growing con-ditions for the grass cover on river dikes (Sýkora & Liebrand, 1987; van der Zee, 1992). The next step was to test the feasibility of the ecological engineering measures proposed in the above mentioned projects empirically.In the research project described in this thesis the core questions were therefore whether the valuable, species-rich vegetation on the dikes can return after reinforcement works, and, if so, what are the pre-conditions for this during and after the reinforcement. The research was carried out on the basis of data of 209 permanent quadrats divided over 125 trial fields. Each permanent quadrat has its own specific method of reconstruction, sowing and management.Ninety-eight percent of the plant species found before reconstruction, reappeared after recon-struction. Most species reappeared on the replaced former top layer. Only a few (rare) species did not reappear but were still present in the unmodified zone. Most relatively rare species occur only in low numbers and consequently they are still at risk of disappearing, especially if no proper management is applied. Because of this, a spared zone seems to be the best guarantee for the conservation of the plant species after the reconstruction.In the period 1987-1994 9 plant communities were distinguished within the vegetation of the experimental river dike. They can be classified as follows on the basis of method of reconstruction, management and successional stage: the Arrhenatheretum with Peucedanum carvifolia and Rumex thyrsiflorus (I) is typical for the spared zone, the species-poor Arrhenatheretum with Urtica dioica and Valeriana officinalis (II) is a rough vegetation resulting from bad management, the Arrhenatheretum with Leucanthemum vulgare and Lysimachia nummularia (III) is a species-rich grassland occurring under good and moderate management practices on replaced sods and replaced topsoil, the Arrhenatheretum with dominance of Alopecurus pratensis (V) is an intermediate vegetation which will develop further, either into a hayfield vegetation or into a pasture vegetation, depending on the management applied, the Lolio-Cynosuretum with Crepis capillaris and Ranunculus repens (VI) is a grassland vegetation strongly influenced by grazing and the Arrhenatheretum with Leucanthemum vulgare and Trifolium pratense (IV), the association fragment of the Arrhenatheretum with Phleum pratense and Ranunculus repens (VII), the fragmentary community with Matricaria maritima and Plantago major [ Arrhenatherion/Chenopodion ] (VIII) and the fragmentary community with Capsella bursa-pastoris and Poa annua [ Eu-Polygono-Chenopodion ] (IX) are pioneer stages, which had wholly or almost disappeared in 1994.The best way to assure maintenance of species-rich grassland vegetation on reconstructed river dikes is to spare a strip or zone of this vegetation during the reconstruction. Species disperse from here to other parts of the dike and the redevelopment of the vegetation is stimulated. To ensure opti-mal results, the soil composition of those new parts should resemble the soil composition of the spared zone as much as possible. If it is not possible to save part of the original vegetation, the upper soil layer can be set aside in the form of turves or as topsoil and can be replaced as the new topsoil after the reconstruction. Replacing the original topsoil after the reinforcement provides a topsoil of similar composition to that before the reinforcement. Besides, the redevelopment of species-rich grasslands is promoted by previously occurring species re-establishing from the propagules present in the replaced topsoil. The application of the under layer as the new top layer and the use of imported clay as the new top layer both prevent a quick restoration of botanically valuable, semi-natural, species-rich grasslands. Propagules are very rare or even absent.The seed mixtures applied influence the development of succession. Application of seed mixtures gathered locally accelerates succession. Seed mixtures containing a considerable proportion of Lolium perenne seeds are unsuitable, as the redevelopment is retarded, especially when applied in the high densities (such as 70 kg.ha <sup>-1</SUP>) which used to be common. Sowing an annual grass species like Lolium multiflorum or the standard seed mixture D1 in a low density of 20 to 25 kg.ha <sup>-1</SUP>seemed not to retard the development of a species-rich vegetation.In the first years after reconstruction the influence of the methods of reconstruction and the seed mixtures applied appears to be preponderant. In the first years the structure of the vegetation is quite open and the competition between species is low. When the vegetation closes, competition increases. Subsequently, management of the vegetation can be used as an important means to regulate competition and, consequently, species composition. A species-rich vegetation only develops when managed properly.On the basis of erosion resistance features like openness of the sward, ground cover, root density and shear resistance, the best management practices appear to be grazing in June in combination with hay-making in September, hay-making in June in combination with grazing in September and hay-making twice a year. In this respect, grazing twice a year, grazing during the whole season, hay-making in September and hay-making in June in combination with mulching in September are moderately effective. Hay-making in June, mulching twice a year, hay-making once every two year, burning and no management are bad management practices.On the basis of ecological features such as species-richness and number and proportion of rare species, the best management is hay-making twice a year. In this respect, hay-making in June in combination with mulching in September, hay-making in June, hay-making in September and hay-making in June in combination with grazing in September are moderately effective. The other grazing practices, mulching twice a year, hay-making once every two year, burning and no management are bad management practices