Keith Beven, Professor, Lancaster University, UK and Uppsalla University, Sweden
There are many different studies in many different countries in which future projections of climate change have been used as inputs to hydrological models as a way of estimating the future hydrological impacts of change. This is being funded and done to inform decision making about how society should react to future change. But the future climate projections are uncertain and do not provide good reproduction of precipitation for the historical control period, nor of the last decade. And the hydrological models are uncertain.
It follows that the "evidence" might not be robust and the evidence might therefore have an effect on what decision is made. There are risk based decision making frameworks available but they generally require that the probabilities be considered complete, and that the loss function can be quantified. The epistemic nature of many of the sources of uncertainty in this approach suggests however that a consideration of the investment required to protect against a given level of change might be a more realistic approach that does not depend on climate projections. How much to invest then becomes as a political decision as to how risk averse (or accepting) a society (government) wants to be.