Populations of marine organisms on coral reef islands (CRI) are connected in space and time by seawater that transports propagules of plants, animals, and algae. Yet, despite this reality, it is often assumed that routine replenishment of populations of marine organisms on CRI is supported by locally-sourced propagules (hereafter, larvae). Following large disturbances, however, distantly-sourced larvae from less disturbed CRI within a regional meta-population are likely to be important for local population recovery, but evaluating the roles of locally- versus distantly- sourced larvae remains difficult. While larval sources are relatively well known for many fishes, they remain virtually unknown for most taxa, particularly those associated with the benthos, including hermatypic corals. We make the case that CRI provide natural laboratories in which studies of connectivity can enhance understanding of community dynamics under future disturbance regimes, especially where ongoing changes have created novel systems that are functioning in ways differing from the recent past. However, this potential cannot be realized due to the limited breadth, detail, and spatio-temporal concordance of exiting research. Targeted research on the role of connectivity in mediating ecosystem resilience of CRI is required to understand how populations of marine organisms will change in a future affected by large-scale disturbances of anthropogenic origin. Using the coral reefs of Mo’orea (French Polynesia), Okinawa (Japan), and St. John (US Virgin Islands) as examples, we describe the data required to achieve this objective, and discuss why provision of these data will require new modes of multidisciplinary and collaborative research.