Towards a better understanding of potential impacts of climate change on marine species distribution: a multiscale modelling approach

Aim In this paper, we applied the concept of ‘hierarchical filters’ in community ecology to model marine species distribution at nested spatial scales. Location Global, M editerranean S ea and the G ulf of G abes ( T unisia). Methods We combined the predictions of bioclimatic envelope models ( BEMs ) and habitat models to assess the current distribution of 20 exploited marine species in the G ulf of G abes. BEMs were first built at a global extent to account for the full range of climatic conditions encountered by a given species. Habitat models were then built using fine‐grained habitat variables at the scale of the G ulf of G abes. We also used this hierarchical filtering approach to project the future distribution of these species under both climate change (the A 2 scenario implemented with the M editerranean climatic model NEMOMED 8) and habitat loss (the loss of P osidonia oceanica meadows) scenarios. Results The hierarchical filtering approach predicted current species geographical ranges to be on average 56% smaller than those predicted using the BEMs alone. This pattern was also observed under the climate change scenario. Combining the habitat loss and climate change scenarios indicated that the magnitude of range shifts due to climate change was larger than from the loss of P . oceanica meadows. Main conclusions Our findings emphasize that BEMs may overestimate current and future ranges of marine species if species–habitat relationships are not also considered. A hierarchical filtering approach that accounts for fine‐grained habitat variables limits the uncertainty associated with model‐based recommendations, thus ensuring their outputs remain applicable within the context of marine resource management.