PREDICTION OF HIGH POTENTIAL AREAS OF HABITAT FOR MONITORED SPECIES

摘要: AbstractThe area of distribution a species is one the fundamental expressions itsecology and evolutionary history. Detailed knowledge areas relevantto address basic questions in biogeography ecology management ofbiodiversity for conservation or sustainable use. The problem we tackle this paperis inferring zones high potential habitat under study, basedon reported sites presence, where each site associated with values covariates,measured on discrete scales. We compute predictive probability that speciesis present at site, by means mixture involving all pairs covariates. Possiblespatial bias presence accounted for. Since posterior distributiondoes not have closed form, MCMC implemented. However, also describe anapproximation to distribution, which avoids MCMC. Available aprioriinformation regarding can be incorporated aclear-cut way. In addition propose map uncertainty allows greaterinsight into nature distribution. By simulations, compareour approach other standard methods. Two case studies are presented.Keywords: Area Biodiversity, Mixture model, Predictive probabilitymap, priori elicitation.