Thirty years of change in the vegetation communities of three valley mires in Suffolk, England.
Detailed floristic records from 12 mires on 3 Suffolk fens were gathered in 1959 and 1991. The data were analysed using tabulation and ordination to elucidate change at the plant community and individual species levels. Substantial changes at both levels were observed. The respective roles of lack of traditional management, changes in water regime and increased fertility in accounting for these changes were explored. At community level, those mires which showed the greatest alteration were those subject to hydrological disturbance combined with dereliction. In terms of the British National Vegetation Classification they changed from short sedge mires allied to M13 Schoeno-Juncetum towards highly fertile tall herb fens with many ruderals and allied to S26 Phragmites australis-Urtica dioica fen. Partial dewatering led to M27 Filipendula ulmaria-Angelica sylvestris tall herb fen with few non-fen ruderals. Mires where ground water influx was reduced or eliminated but where traditional management was sustained showed less community change, with more of the principle community dominants being maintained. Here M13 Schoeno-Juncetum mires developed into dry variants of M24 Molinia caerulea-Cirsium dissectum fen meadows. Mires with undisturbed hydrology but which became derelict were transformed from M13 Schoeno-Juncetum into hydrophytic Cladium-dominated communities of the S25 Phragmites australis-Eupatorium cannabinum tall herb fens. Many fen species were retained or recruited, while the colonization of ruderals, scrub and non-fen species that was noted in the dewatered mires did not occur. The dewatered mires showed the greatest alteration at species level. In unmanaged areas, all of the fen associates and bryophytes were lost and replaced by non-fen species and ruderal tall herbs. Where traditional management was maintained, most of the small species characteristic of wet fens and semi-aquatic conditions were still lost, although recruitment of ruderals and non-fen plants was less evident, being restricted to species of mesotrophic meadows. Retained species tended to be common fen plants or non-fen species. Mires which were derelict, but where the natural hydrology persisted, showed the least change in associated flora. Nevertheless, many of the smaller herbs were still eliminated by competition in the now dense tall-herb fen, indicating that traditional management was required as well as hydrological integrity for the survival of most wet-fen species. However, some wet-fen species were retained or recruited, indicating management to be subordinate to hydrology in their maintenance. They included Caltha palustris, Valeriana dioica, Lotus uliginosus and the bryophytes Campylium stellatum, Aneura pinguis, Sphagnum subnitens and Riccardia multifida. All mires showed a decline in conservation importance as measured by indices of botanical 'quality'. The mires which showed least decline (15-64% reduction in Rarity Weighted Principal Fen Species Score) were those which retained their hydrological integrity, regardless of management status. One mire, only partially dewatered but still managed, declined by nearly 70%. Tentative comments are made on the relative importance of the 3 environmental factors to the maintenance of mire systems. It is concluded that maintenance of stable hydrologies is more important than extensive management in the long-term conservation of these sites.