Meta-ecosystems are the largest and probably most complex structures investigated in ecol-ogy. Because of their complexity they are often separated into their respective ecosystems and then studied in isolation. This is often done without analysing whether an understanding of the individual ecosystems can lead to a proper understanding of the meta-ecosystem, like-ly because we lack clear guidelines when such a separation is sufficient. We here propose four conceptual models for which a separation appropriately approximates the full dynamics. For each of these models we provide empirical evidence from riparian meta-ecosystems to showcase that the terrestrial or aquatic ecosystem should likely not be studied in isolation. Finally, we provide a new theoretical framework to assess how strongly two ecosystems are linked and discuss how this framework might be applied in future empirical research.