7. Concluding remarks
Present comprehension of climate change and its past, present, and future impacts on water resources generally suffers from strong uncertainties because of the gaps in knowledge and insufficient understanding of the complex processes and their feedbacks and interconnections (cf. Funtowicz and Ravetz,1990).
Even if the stock of knowledge increases with time, we find factors that were ignored earlier, hence the uncertainty may also grow. It also seems to grow if we use multi-model ensembles that lead to different simulations, instead of single models.
Remarkable uncertainty of climatic input comes on top of the “traditional” uncertainty existing in hydrology, related to portraying links between variables, choice of model structure, parameterization and parameter estimation. We reviewed metrics for representing the degree of uncertainty and demonstrated how uncertainty can be reduced.
There are many sources of uncertainty in projections for the future, discussed in this paper. Using ensembles of climatic and hydrological models may allow to find more robust results, supported by several models, for some river basins or regions. But, at the same time it can result in large spread of simulations in other regions, and even different directions of change that are difficult to interpret using current tools.
There exists an opinion that decision making should be postponed until sufficient knowledge becomes available (wait-and-see stance). However, as noted by Refsgaard et al. (2013), in some contexts even large uncertainties may imply small consequences for decision making – there can be sufficient knowledge to justify action in climate adaptation.
