Modelling dispersal in a large parrot: a comparison of landscape resistance models with population genetics and vocal dialect patterns

Context Identifying the range, core areas and dispersal pathways or barriers in heterogeneous landscapes is important for managing threatened species. Studies of variation in learned vocalisations are a promising complementary tool to traditional landscape genetics studies for identifying potential dispersal barriers. Here we use multiple data sources to inform the conservation of a parrot species. Objectives We tested for correlations between landscape resistance models, population genetic structure and vocal variation of parrots to investigate the effects of natural barriers on genetic and behavioural population structure including narrow habitat corridors and a mountain range. Methods We studied palm cockatoos (Probosciger aterrimus) within their Australian distribution. We constructed landscape resistance surfaces restricted to areas of high climatic suitability from a maximum entropy (MAXENT) distribution model. We verified three landscape resistance predictions from CIRCUITSCAPE (isolation by elevation, habitat and distance) using four data sets (individual genetic divergence, acoustic divergence in repertoire and two call types). Results Landscape resistance models revealed strong effects of isolation by elevation on genetic, repertoire and structural call differentiation. Neither isolation by habitat nor isolation by distance were well supported by differentiation in the data. Conclusions Our landscape resistance analysis validated by four datasets supports the Great Dividing Range as the main limitation on dispersal and connectivity among palm cockatoo populations. Combined genetic and behavioural approaches can determine landscape-level connectivity of individuals and demonstrate how dispersal barriers influence genetic and behavioural patterns in a large parrot.This research is supported by an Australian Government Research Training Program Scholarship