Abstract
Marine reserves have increasingly been recognised for their potential to address the pervasive problem of unsustainable harvest of fisheries worldwide. Abalone catch around Stewart Island has been in decline since the turn of the century despite being managed under NZ’s quota system. A nested logit model is applied to spatially recorded catch and effort data between 1998 and 2003 to capture the two level decision making process of divers. On any given day, divers decide whether to go diving at all, and if yes, which of the 16 statistical areas around Stewart Island to visit. Weather conditions, patch-specific levels of catch per unit of effort and distance used as explanatory variables lead to a reasonably good fit, and the model can be used to simulate the imposition of a network of no-take areas as an additional management tool to NZ’s quota system. Suitable areas are selected according to criteria that yield the ‘least economic impact’. The results show that most effort redistributes to adjacent areas, which could potentially outweigh the benefits of no-take areas. This highlights the fact that incorporating fishermen behaviour is crucial in the marine reserve debate.