 The impact of climate change on water resources can be estimated by post-processing regional climate model, RCM, simulations. These simulations are prone to systematic biases in precipitation, so they must be adjusted to better reflect observed climatologies. Three types of transformation were examined, distribution derived, parametric and non-parametric. Non-parametric transformations were found to best reduce biases in RCM precipitation. This was demonstrated through a case study of 82 Norwegian precipitation stations. This article was authored by J. E. Haugen, T. Engenskogen, J. B. Brems and others.