Korapuki Island as a case study for restoration of insular ecosystems in New Zealand

Aim Success with eradicating invasive species from islands around New Zealand raises the prospect of reversing the loss of species by restoring biotic communities on modified islands. I seek to identify methods that can be used to clarify restoration targets on Korapuki Island, which was modified by introduced mammals until 1987. Location Korapuki, Green and Middle islands, Mercury Islands, north-eastern New Zealand. Methods I describe methods of identifying restoration targets by using biogeographical frameworks, and benchmark systems based on neighbouring unmodified islands. Methods that can help identify directions and rates of change are also investigated. Results The benchmark islands (Middle and Green) have complex forest systems with a distinctive component of large, flightless invertebrates, dense populations of tuatara Sphenodon punctatus (Gray), dense and diverse populations of lizards and dense populations of burrowing seabirds. The restoration island (Korapuki) has simplified forest systems, few large invertebrates, no S. punctatus, a depauperate fauna of lizards and scattered populations of seabirds. These differences are consistent with the recorded effects of rabbits Oryctolagus cuniculus (L.) and Pacific rats Rattus exulans (Peale) removed from Korapuki Island in 1986-87. Three restoration scenarios are described and restoration targets identified. Time scales for natural succession are likely to be influenced by a long-lived colonizing tree species that now predominate over much of the island. The rate of expansion of reintroduced animals is influenced by the very low productivity of many of the species apparently extirpated by introduced mammals. Main conclusions Ecological restoration of island ecosystems has been likened to reconstituting the ambiguous because of conceptual and practical difficulties. Goals for restoration of island systems are often dependent on value judgements. Biological outcomes or targets can be clarified by the use of unmodified neighbouring islands as benchmarks. However, successional pathways on the restored island may not converge with the benchmarks because of environmental differences between sites.