10276: Discrimination of semi-natural plant communities from abandoned fields by ordination and neural networks

Ejrnæs, Rasmus; Liira, Jaan and Poulsen, Roar S. (2003) Discrimination of semi-natural plant communities from abandoned fields by ordination and neural networks. [preprint]^*

Summary

This study deals with the succession from abandoned fields to semi-natural grassland and heathland vegetation and the discrimination between these types of habitat. We hypothesise that semi-natural condition may be indicated by species lists of vascular plants. A statistical classification model is developed, based on 2059 reference samples from Danish ancient grasslands and heathlands, and abandoned fields of varying age. This Succession Model is shown to discriminate effectively between abandoned fields and semi-natural habitats, and it is suggested to be useful for the detection of conservation-worthy abandoned fields. A test of four hypotheses regarding the model prediction of naturalness of abandoned fields revealed that successional age, period of abandonment and successional trajectory had significant impact on the succession on abandoned fields. The implications of the results for restoration of semi-natural habitats and the usefulness of the model in practical conservation management are discussed.

Archive Staff Only: edit this record