ELSEVIER Biological Conservation 73 (1995) 151- 160 © 1995 Elsevier Science Limited Printed in Great Britain. All rights reserved 0006-3207/95/$09.50+ .00 0006-3207(95)00052-6 MODELLING THE REINTRODUCTION OF THE GREATER BILBY Macrotis lagotis USING THE METAPOPULATION MODEL ANALYSIS OF THE LIKELIHOOD OF EXTINCTION (ALEX) Richard Southgate Conservation Commission of the Northern Territory, PO Box 1046, Alice Springs, NT 0871, Australia & Hugh Possingham Department of Applied Mathematics, University of Adelaide, GPO Box 498, Adelaide, SA 5001, Australia (Received 16 January 1994; accepted 17 May 1994) Abstract Population viability analysis of the greater bilby Macrotis lagotis was conducted using the metapopulation model ALEX. ALEX was used to examine the effect of reintro- duction configurations and potential manipulations on the extinction risk of the metapopulation. More specifically, the impact of patch separation, sporadic big rains and fire on population survival was investigated. Estimates of population parameters, including survival and fecundity, dispersal rate and home range size, were derived from a bilby reintroduction programme con- ducted in central Australia, as were parameters describ- ing environmental quality and the impact of fire, drought and sporadic large rainfa,!l events. The modelling results indicated that local segregation of a population to achieve replication did not change extinc- tion risk dramatically and that an experiment for a 2-year duration (equivalent to a generation time) could be achieved at low risk with a small population. Adult mor- tality was of key importance in determining population survival. With the release of 12 individuals, drought and sporadic high rainfall events limited the life of the reintro- duced population to a median of 8 years. In the absence of these events, median extinction time was extended to 26 years. The modelling indicated that there was little scope to reduce environmental variation (and improve the popula- tion survival prospects) through manipulation or manage- ment of the environment. However, spatially subdividing subpopulations so that they should experience appreciably different big rainfall and drought events has potential to reduce metapopulation ex,!inction risk. Keywords: Population viability analysis, reintroduction, bilby. INTRODUCTION The application of a population viability analysis model in the formulation of a species reintroduction 151 programme is discussed using the greater bilby Macro- tis lagotis as an example. The bilby is a type of a bandi- coot. It is nocturnally active, semi-fossorial, constructing a deep burrow for shelter, and omnivo- rous consuming predominantly seed, bulbs and inverte- brates (Johnson, 1989). The species is strongly sexually dimorphic in size with adult males weighing 1500-2500 g and females 700-1100 g. It was once distributed over 70% of the arid and semi-arid parts of Australia but declined rapidly in the late 1800s and early 1900s and is now considered vulnerable to extinction by the IUCN (Kennedy, 1992). The remnant populations are restricted to relatively unproductive areas of the former range where stock, rabbits Oryctolagus cuniculus and foxes Vulpes vulpes are absent or rare (Southgate, 1990a). A programme was initiated in the mid 1980s to determine if reintroduction could be used to expand the species' current range. The role of reintroduction in conservation is gener- ally perceived to be one of providing a management tool for the creation of additional subpopulations or to bolster an existing population. This notion has been reinforced in a policy produced by the Reintroduction Specialist Group of the IUCN (1987), which advocates that reintroduction be used only when the limiting pro- cesses operating at a release site have been eliminated. This presupposes that original causes of decline can be clearly determined and that favourable habitat can be identified and managed effectively. However, it is rare for the habitat requirements of a threatened species to be thoroughly understood and frequently the threaten- ing processes are not clearly evident (Scott & Carpen- ter, 1987; Grittith et al., 1989; Short et al. 1992). Under these circumstances, the acceptability of habitat condi- tions must be investigated by monitoring the survival, fecundity and dispersal characteristics of the species. In other words, reintroduction must be used as a research tool to define habitat conditions and continue in that role until habitat favourability can be predicted with a