Spatial management measures in the form of no-take areas used in fisheries management can provide a buffer against catastrophic events. Dynamic area closures, like rotational closures, have also been used as a management tool particularly for sessile organisms. In this study, bioeconomic models are developed to investigate dynamic closure strategies for use as a management tool in the harvest of a metapopulation consisting of two local sub-populations. The models provide an optimal strategy that maximises the sum of discounted net returns with a fixed harvest level [i.e. total allowable catch (TAC)] by opening and closing the sub-populations of a metapopulation, subject to random negative catastrophic effects. Results showed the optimal policy for opening and closing a single exploited population depends on the degree and pattern of migration between it and other sub-populations. When the harvest or TAC can be applied to either sub-population, the optimal closure strategy depends on the abundance of both populations, crucially, even if they are biologically independent. The results provide insights into the management of stochastically fluctuating populations including more mobile species that are frequently not subject to no-take controls.