The classical metapopulation theory and the real, natural world: a critical appraisal

Classical metapopulation (CM) theory considers that species persistence in the landscape depends on a turnover of extinction-(re)colonisation of suitable habitat patches at each generation. Therefore, metapoputation dynamics are approximated by binary changes in the state of individual patches. This approach has allowed rigorous mathematical analyses of metapopulation dynamics, establishing links with general laws in ecology or deriving operational concepts in conservation biology. However, a critical review of the literature shows that only a few spatially structured populations function according to CM theory. Moreover, empirical studies of CMs usually focus on species (1) at the margin of their distribution range, (2) with small or extremely small local population sizes and (3) declining in the study area (expanding in one case). Such non-equilibrium population systems are far from the equilibrium world idealised in CM models. Furthermore, empirical comparisons of metapopulations in highly fragmented landscapes show that extinction-(re)colonisation dynamics match those predicted by the theory only when local populations are very small. The lack of correspondence between CM theory and the real, natural world may explain why models and concepts derived from the theory are not used in population viability analysis or conservation strategies. The final message of this paper is three-fold: (1) exclusive, implicit assimilation of spatially structured populations to CMs should be avoided, (2) attempts to manage any population using CM theory should be preceded by careful examination of the correspondence between the field situation and assumptions of the theory (particularly regarding the high rate of local population turnover and the equilibrium state of the metapopulation) and (3) theorists should avoid to present CM theory and derived concepts as the only conservation framework for populations confronted with habitat loss and fragmentation. (C) 2004 Elsevier GmbH. All rights reserved.