A climate change scenario for the Nordic countries has been defined for application in hydrological models in the Nordic research project 'Climate Change and Energy Production'. The scenario is based on a subjective evaluation of several recent results from global coupled atmosphere and ocean general circulation models (GCMs) and on a statistical downscaling of the model results. The scenario specifies a warming rate which ranges from 0.3 degrees C per decade for Iceland and the Faeroe Islands to 0.45 degrees C per decade for eastern Finland and the northernmost parts of Sweden. There are marked seasonal and regional differences in the warming rate. Summer warming is relatively uniform over the area, ranging from 0.25 degrees C per decade in Iceland and the Faeroe Islands to 0.3 degrees C per decade in Finland. Winter warming is more variable, ranging from 0.35 degrees C per decade in the North Atlantic area to 0.6 degrees C per decade in Finland. Precipitation is increased by 3 to 6% per degree of warming, yielding an increase ranging from 1% per decade during the summer season in Finland, Sweden and eastern Norway to 2.5% per decade during the winter season in western Norway. Climate in the Nordic countries is characterized by low-frequency natural variability on decadal time scales which is thought to be partly driven by changes in the thermohaline circulation of the North Atlantic. Future behaviour of the North Atlantic ocean circulation is highly uncertain, and model predictions of climate development in this part of the world are perhaps more difficult than for most other regions. Other sources of uncertainty in GCM model computations apply to the North Atlantic region as well as to other regions of the world, e.g. coarse model resolution, somewhat arbitrary coupling of the ocean and the atmosphere and uncertain parameterizations of clouds and precipitation. In view of these difficulties, the scenario must be considered an extremely uncertain, although plausible, description of what might happen in the future rather than a prediction of what most likely will happen.
