Succession during the re-creation of a flood-meadow 1985-1999

The study site, Somerford. Mead, is located on the river Thames floodplain and was a species-rich flood-meadow in the 1950s, In the 1960s and 1970s it was subjected to intensive grassland management with regular NPK additions and occasional herbicide treatment. In 1981 Somerford Mead was ploughed for the first time and converted to arable land. Seeds of an Alopecurus pratensis-Sanguisorba officinalis flood-meadow community MG4; Rodwell 1992) were sown onto prepared soil in the autumn of 1986, and botanical records were made from 1985 to 1999. From 1989 to 1999, three replicates of three treatments: cow-grazing, sheep-grazing and no-grazing were introduced after hay-cutting. Analysis successfully separated the establishment phase from the experimental phase and showed a significant difference between the grazed and ungrazed treatments. Abiotic and biotic factors which might contribute to successional trends are discussed. A convoluted pattern for each treatment could be attributed in part to intrinsic 'cycles' of perennial hemicryptophytes behaving as short-lived species and in part to the percentage frequency of many species which was reduced in 1990 and 1995/1996, years of drought. After the initial inoculation of MG4 seed and the disappearance of arable therophytes, recruitment of new species was very slow. Coefficients for Somerford Mead matched against MG4 (Rodwell 1992) produced an equilibrium within three years. It subsequently fluctuated over a 10-yr period well below the level of Oxey Mead, the donor site. Land managers should ensure that their proposed site has the right soils and hydrology for MG4 grassland and that traditional management of hay-cutting and aftermath grazing is practised. Only one cut a year in July could lead to a reduction in percentage frequency of most species except Arrhenatherum clatius.