Both movement ecology, which focuses on modelling patterns of organismal movement across landscapes, and metapopulation and metacommunity ecology, which focuses on how patterns of dispersal across landscapes shape communities, have been the focus of growing ecological research in the last decade. However, there has, to date, been very little research into how to derive the structure of a metapopulation model (the patches and connections between them) from modern movement models. In this talk, I will lay out the ambiguities in metapopulation ecology about what a patch is that make this task difficult, and will discuss a promising approach to solve this: defining what a patch is based on residency time, and using modern graph-based statistical clustering methods applied to a given movement model to divide a landscape into patches.