Dr Rachel McCrea - University of Kent
Fri 07 Dec 2018, 14:05 - 15:00
Bayes Centre, Room 5.46

If you have a question about this talk, please contact: Ruben Amoros Salvador (ramoros)

Image for Maximising the potential of removal data

The seminar will start one hour before the usual time.

Joint work with Ming Zhou, Eleni Matechou and Diana Cole.

Removal surveys typically involve daily visits to a site with a record being kept of the number of animals being collected on each sampling occasion. The recorded data are of the form of counts, where nt denotes the number of animals collected and removed on sampling occasion t. Under the assumption of closure and constant capture probability the expected number of captures at each sampling occasion follows a geometric distribution and enables the estimation of the unknown population size, N.  Within this talk I will present three new approaches for dealing with violations of the assumption of population closure. The first allows for new arrivals to enter the population with the unknown number of new arrival groups selected using a reversible jump MCMC algorithm; secondly I will present a new approach which implements the LASSO to determine when new arrivals enter the population and finally I will present a multi-state model with an unobservable state to account for temporary emigration which can be fitted to data collected under a robust design sampling scheme.