a, A framework for evaluating risk of exposure occurring by the year 2050, when global mean sea-level rise is similar under different emission scenarios, based on the likelihood of hazards related to sea-level change occurring in any given area during a given year. Mean sea level is indicated by the border between blue and orange regions. b, Exposure and migration dynamics that may emerge simultaneously in this period, their various triggering mechanisms and potential policies that may act as intervening factors. Migration is depicted (arrows) between potential destinations (red circles, the size of which correspond to rising and falling rates of migration). In the short term, policy may simultaneously influence various forms of mobility and migration at sea level including (A) short-distance, temporary or seasonal migration, (B) urbanization in coastal cities, (C) permanent migration from rural areas, (D) mobility traps, (E) prospective migration in advance of impacts, (F) short-term ‘climate gentrification’ of high-risk areas and (G) background pro-coastal migration. c, The possibility space for one projection of sea-level change under an unspecified emission scenario at the year 2100, with associated adaptation interventions and migration outcomes. Areas in flood plains will be inundated with increasing frequency (dark orange), and new areas formerly above the annual flood plain will experience intensifying risk (orange). Adaptation interventions will increasingly affect migration dynamics, including (A) the concentration of multiple migration flows to fewer safe destinations where coastal fortification and accommodation has occurred, (B) settlement abandonment where no adaptation interventions have occurred, (C) community relocation accompanying a strategy of retreat and (D) sudden migration as adaptation interventions reach their limits.