If impact studies are available, the next decision node to consider is whether these studies are credible and have comprehensively explored the full range of uncertainty. Impact projection is only useful for adaptation if a representative range of uncertainties in terms of climate and socio-economic scenarios is explored because adaptation decisions that are based on only a limited range of scenarios may lead to maladaptation. Ideally, it would even be desirable to use a range of impact models for these projections. In practice, however, impact models are only available for some sectors such as agriculture, forestry and water and even then uncertainty in impact models are often large (Hoffmann et al. 2011). Most impact assessments found in the literature, however, only explore a subset of this uncertainty range (Hofmann et al 2011; Hinkel 2011b).

Further criteria apply to the credibility of the impact models. Are the available models well calibrated on a robust empirical basis? Impact models themselves are uncertain, and thus ideally impact projection should also make use of several impact models in order to characterise uncertainty. These issues are discussed in greater depth in the Toolbox section on Modelling future impacts.

If existing studies are not credible and/or comprehensive, then further impact analysis is indicated.