Thresholds in climate migration

Migration in response to climatic hazards or changes in climatic conditions can unfold in a variety of ways, ranging from barely observable, incremental changes in pre-existing migration flows to abrupt, non-linear population movements. The adoption of migration instead of in situ adaptation responses, and the high degree of variability in potential migration outcomes, in part reflects the presence of thresholds or tipping points within the processes of human-environment interaction through which climate adaptation and migration take place. This article reviews and makes linkages between existing research in climate adaptation, migration system dynamics, residential preferences, and risk perception to identify and explore the functioning and importance of thresholds. Parochial examples from the author's published research on climate adaptation and migration in rural North America are used to illustrate. Six types of thresholds in response to climate hazards are identified: (1) Adaptation becomes necessary; (2) Adaptation becomes ineffective; (3) Substantive changes in land use/livelihoods become necessary; (4) In situ adaptation fails, migration ensues; (5) Migration rates become non-linear; and (6) Migration rates cease to be non-linear. Movement across thresholds is driven by context-specific characteristics of climate events, natural systems, and/or human systems. Transition from incremental to non-linear migration can be accelerated by people's perceptions, by actions of influential individuals or groups, and by changes in key infrastructure, services, or other community assets. Non-linear climate migration events already occur at local and sub-regional scales. The potential for global scale, non-linear population movements later this century depends heavily on future greenhouse gas emission trends. The ability to identify and avoid thresholds that tip climate migration into a non-linear state will be of growing concern to policy makers and planners at all levels in coming decades. This article forms part of a special issue of this journal dedicated to the late Graeme Hugo, and the author draws heavily on past research by Professor Hugo and colleagues.