Should metapopulation restoration strategies increase patch area or number of patches?

Managers of species that exist as metapopulations are faced with many decisions. In this paper we use a decision-theory framework to examine a fundamental management question: Should we focus on decreasing the local extinction probability of subpopulations by increasing the size of their patch, or should metapopulation viability be improved by constructing more patches? Using a spatially implicit stochastic metapopulation model and stochastic dynamic programming (SDP), we found the optimal solution to this problem for both the finite- and infinite-time horizon cases. We showed that the SDP solutions outperform a range of heuristic management strategies. The optimal strategy for a given parameter set depends heavily on metapopulation parameters, and it is difficult to make generalizations about the optimal restoration strategy a priori. Although heuristic strategies perform well in some cases, it is not possible to judge their performance until the SDP solution has been computed, and for this reason we advocate the use of SDP as a management tool in restoration. We demonstrate the use of SDP by deriving an optimal management strategy for a population of the Mount Lofty Ranges Southern Emu-wren (Stipiturus malachurus intermedius).