Due to uncertainty about mid to long term impacts will continue to make the construction of probability density functions for impacts problematic (Adger et al., 2009). Several authors argue that for this reason optimal adaptation decision making should be abandoned in favour of robust decision making for adaptation. Robust-decision making requires running a large amount of scenarios (without probability attached) and analysing alternatives over these scenarios on a given set of criteria. It does not require probability functions to the different scenarios. This way options can be eliminated which do not perform well in projected futures, even when the likelihoods of future evolutions are not well known. For example, Wilby and Dessai (2010) apply a method of robust decision-making to address the question of ranking adaptation options in the water sector in Wales and the UK. The method identifies options that address policy goals in the current climate, then tests the sensitivity of the costs and benefits of these options to a large number of future scenarios. Due to high uncertainty in climate models at the scales needed for adaptation decisions, a cost-benefit framework is used to identify robust options, where the benefits exceed costs across a wide range of scenarios of future impacts of climate change. Those measures that have a negative benefit-cost ratio for some projected future climate are not considered robust. They argue that measures that are flexible and permit updating according to future conditions are more likely to be robust to future climate changes; though there may be other robust options that are not flexible.