Climate model precipitation in hydrological impact studies: limitations and possibilities / Nederbörd från klimatmodeller i hydrologiska effektstudier: begränsningar och möjligheter
Hydrology is strongly precipitation-dependent and hence hydrological modeling and forecasting requires accurate precipitation input for good performance. Hydrological climate change impact assessment thus requires good estimates of future precipitation. Climate modeling is the main means of estimating future precipitation, but it is seldom possible to directly use the precipitation output from climate models for meaningful hydrological simulation. This is because (i) climate model precipitation is generally biased, i.e. it deviates from observations in a historical reference period, and (ii) the model grid cells are often far larger than catchments, which creates scale effects. In this paper we give an overview of how climate model precipitation differs from observations in Sweden, at different scales. Then two approaches to bias correction and downscaling that has been developed and applied within HYDROIMPACTS2.0 are described: Delta Change and Distribution-Based Scaling. We close the paper with some reflections on on-going and future research directions.