Abstract
We agree with Costello and Connor [1] that most marine reserves should not be considered as totally isolated islands. However, the marine environment is not a homogeneous habitat either. Landscape connectivity quantifies the extent to which the movements of genes, propagules, and individuals between populations (i.e., dispersal) are facilitated or hindered by the structure and composition of the landscape [2]. Landscape [3] and seascape [4,5] genetics in particular provide a framework (concepts, tools, community) that considers explicitly the suitability of habitats outside marine reserves for the estimation of genetic connectivity [6]. Yet, although tools are available (Box 1), the vast majority of studies that estimate connectivity in the context of marine conservation and strategic management planning still ignore seascape resistance models (Box 1). In this respect, it is true that marine reserves in particular, and marine populations more generally, are often envisioned as islands. We advocate for sampling designs that are more continuous across the landscape, including areas and populations inside and outside marines reserves but also in between .