The difficulty of making predictions of future patterns of climate change migration springs principally from the fact that, methodologically, it is very difficult if not impossible to unpack the different environmental drivers and triggers of migration. People move for complex sets of reasons of which a changing environment is only one. As such it is important to try to understand the impacts of a variety of factors including social dynamics, institutional capacities, demographic growth, inter-community tensions, social cohesion, natural resource management, poverty, politics and power.
In terms of broader trends and patterns which might be identified, the empirical research does not support the claim that climate change will trigger waves of South-North interregional migration. Most displacements are likely to be within country borders. Cross-border migration, where it does exist, is likely to happen within existing social networks, especially in sub-Saharan Africa.
Longer term changes in conditions may result in longer term displacements. Initially the bulk of population movements is expected to be disaster driven. These may become longer term in situations where either the environment is no longer able to promote return or where other vulnerabilities (extreme poverty, social trends etc) make return impossible.
A commonly held view is that those aspects of climate variability most likely to affect migration are the ones which impact on environmentally based livelihoods and are location-specific. Another point often made in the literature is that those most vulnerable to both climate change and forced migration have the least resources to undertake long-distance migration, so are often left behind.
The extent to which new migratory trends are likely to emerge as a result of climate change is still unclear. Some argue that, as a ‘threat-multiplier’, climate change is more likely to exacerbate existing problems although it is unclear is by how much. Others argue that new trends are likely to emerge (e.g Alaska, Pacific islands) but there is currently only limited case study evidence for this.