Spatial and temporal analysis of vegetation mosaics for conservation: poor fen communities in a Cornish valley mire

Aim Biogeographers increasingly realize the importance of seeing plant communities as spatial mosaics and understanding the spatial and temporal heterogeneity of a site is often a key to successful conservation. The aim of this paper is to examine the approaches to the description and analysis of spatial and temporal variation in sub-communities within patch mosaics of vegetation in order to inform conservation management. The activities of the tin streaming industry in Cornwall over the last century have created a highly varied mosaic of poor fen vegetation on Goss Moor National Nature Reserve (NNR). The wetland mosaics comprise dry hummocks and different sized wet pools. The size and depth of the pools determines the rate and type of vegetation that develops, as does the nature of boundary or edge. The ergodic hypothesis is used to describe the various plant sub-communities and their boundaries to identify pathways of hydroseral succession. A further aim was to test the use of Ellenberg Indicator (EI) values as a tool for the rapid description of spatial and temporal environmental change on wetland sites with a view to their management. Location Goss Moor National Nature Reserve, Cornwall, UK. Methods An extensive survey of the whole wetland complex was undertaken to identify patches of poor fen vegetation containing Potentilla palustris (L.) Scop. and Menyanthes trifoliata L. At each patch, species abundance data were collected as well as associated environmental information such as depth of the organic layer and standing water depth, patch location, patch size and boundary type. The plant sub-communities present were defined using techniques of numerical classification [two-way indicator species analysis (TWINSPAN)] and ordination [detrended correspondence analysis (DCA)] and these were ordered using the ergodic hypothesis in order to characterize the stages of the hydrosere. Floristic and environmental relationships were examined using canonical correspondence analysis (CCA). Further environmental differences between the poor fen sub-community types were characterized by weighted EI values for acidity (R), moisture (F), nitrogen (N) and light (L). Results and conclusions Twelve poor fen sub-community types were described and found to be distributed along a primary environmental gradient of organic matter depth, surface water height and bare substrate. Separation of the poor fen communities by a moisture gradient was considered as spatial evidence for hydroseral succession, which begins with the colonization of open-water pools created by tin excavations. High water levels were associated with the swamp communities, increased organic depth was associated with poor fen, and the type of boundary was shown to affect the resulting community composition. Weighted Community Ellenberg Indicator values for nitrogen, light, reaction and moisture are recommended as an effective tool for indicating differences between plant (sub-)communities. The importance of examining sub- community mosaics in the study of hydroseral development is stressed and the manner in which both sets of information may be used to underpin the conservation management of the site is demonstrated.