Biogeography with jumps, distance and trait-based dispersal Event as iCalendar

(Biodiversity and Biosecurity, Biological Sciences, Seminars)

11 April 2017

1 - 2pm

Venue: Federation of Graduate Women's Room, Old Government House

Location: Corner of Princes Street and Waterloo Quadrant, City Campus

Host: School of Biological Sciences

Cost: Free - all welcome

Contact info: Dr Anna Santure

Contact email: a. santure@auckland.ac.nz

Biogeography with jumps, distance, and trait-based dispersal: Adding realism fits the data better than traditional models in historical biogeography

Dr Nick Matzke, Moritz Lab, Centre for Biodiversity Analysis, Division of Evolution, Ecology, and Genetics, Australian National University

Historical biogeography is the study of how species' geographic ranges have evolved across the globe on evolutionary (phylogenetic) timescales. While probabilistic models are used to make inferences from data in phylogenetics and many other fields, historical biogeography has been slow to join this revolution. 

Much of the difficulty lies in the computational complexity of biogeography models – a DNA transition rate matrix is only 4x4, but a biogeography rate matrix can easily be 1000x1000 or higher, leading to severe computational constraints.  As a result, our answers to fundamental biogeographic questions – for example, the importance of long-distance (‘jump’) dispersal, the relationship between distance and dispersal probability, the relationship between dispersal and key traits – have often been decided ahead of time by simplifying assumptions made in biogeography software, rather than being inferred from the data.  

Dr Nick Matzke wrote the R package ‘BioGeoBEARS’ to overcome some of these difficulties, and to allow more biologically realistic models to be used with geographic range data. Using 20 island and continental clades, Dr Matzke will show that the data usually favours models that allow a non-zero probability of jump dispersal. He will also show that a relationship between dispersal and geographic distance is often supported by the data, and is likely to be a key feature of global historical biogeography analyses. Finally, he will give examples where trait-dependent dispersal ability is inferred from the data.

Despite progress, biogeography is only beginning to be explored with the tools of probabilistic inference. Dr Matzke will outline future research directions, for example, the integration of historical and ecological biogeography with phylogeny-informed species distribution modelling.

Close up portrait image of Dr Nick Matzke against a background of trees
Dr Nick Matzke