The use of probabilistic habitat suitability models for biodiversity action planning

Resources for nature conservation initiatives are generally limited; therefore a spatially explicit targeting approach aimed at identifying the most viable sites has the potential to maximise conservation effectiveness. This research developed a multi-species targeting approach for eight threatened bird species associated with rare neutral grasslands within an agri-environment scheme in the UK. Suitable habitat areas for each bird species were identified through the development of two predictive models. The first 'environmental' model was based on habitat association data developed using multilevel modelling techniques; the second was a geostatistical model developed using indicator kriging techniques. The two models were combined using Bayesian decision rules to produce a refined map of habitat suitability. Probability threshold values symptomatic of sites of highest habitat suitability were chosen in relation to national habitat creation targets, for example the most suitable 500 hectares (ha) for each species were identified. Once determined these areas were overlaid within a Geographical Information System (GIS) to identify sites shared by all target species. As a relatively small area was identified shared by all eight species (1.56 ha) further species assemblages, such as wading bird species, were also investigated. The approach effectively identified land parcels that potentially met the UK biodiversity action plan targets and species requirements of these assemblages where applicable. If applied to other agri-environment schemes this targeting approach could make conservation effort more effective and reliable at both the regional and national scale. (c) 2005 Elsevier B.V. All rights reserved.