Use of Representative Climate Futures in impact and adaptation assessment

A key challenge for climate projection science is to serve the rapidly growing needs of impact and adaptation assessments (hereafter risk assessments) in an environment where there are substantial differences in the regional projections of climate models, an expanding number of potentially relevant climate model results, and a desire amongst many users to limit the number of future climate scenarios in their assessments. While it may be attractive to select a small number of climate models based on their ability to replicate current climate, there is no robust method for doing this. We outline and illustrate a method that addresses this challenge in a different way. The range of plausible future climates simulated by climate models is classified into a small set of Representative Climate Futures (RCFs) and the relative likelihood of these estimated. For each region, the RCFs are then used as a framework in which to classify more detailed information, such as available climate model and downscaled data sets. Researchers wishing to apply the RCFs in risk assessments can then choose to use a subset of RCFs, such as the most likely, high risk and least change cases for their impact system. Preparation and analysis of future climate data sets can therefore be confined to those models whose simulations best represent the selected RCFs. This significantly reduces the number of models involved, and potentially the effort required to undertake the risk assessment. Consistently applied within a region, RCFs, rather than individual climate models, can become the boundary objects which anchor discussion between the climate science and risk assessment communities, simplifying communication. Since the RCF descriptions need not change as new climate model results emerge, they can also provide a stable framework for assimilating risk assessments undertaken at different times with different sets of climate models. Systematic application of this approach requires various challenges to be addressed, such as robustly classifying future regional climates into a small set and estimating likelihoods.