Bayesian belief networks - a potential tool for conservation planning of endangered plant species populations

Abstract

Bayesian belief networks (BBN) have been increasingly used as a potential decision supporting tool useful in conservation management. We assessed the application of the BBN model to support management in conservation planning of Pulsatilla patens (L.) Mill., the endangered plant species on a European scale, as an example. The Bayesian network approach was used to develop a model of the impact of biotic and abiotic variables on the morphological-developmental features and demographic features of the population in NE Poland. Field data collected from the total number of 47 sites in the 4 largest forest complexes were used to develop a model using GeNIe 2.0. The diagnostic testing and sensitivity analysis indicated that the greatest impact on the population features was the number of competing species in the forest undergrowth. Validation has shown that the developed model is effective for evaluation of the impact of habitat conditions on the population features deciding about the reproduction of this taxon. The BBN model was also used to define optimal habitat conditions ensuring regular growth and development of P. patens. Finally, we indicated the protective treatment to help preserving the species considered. Therefore, the developed model is recommended as a potential tool to support decision-making aimed at the conservation planning of endangered plant species.