Global climate change will have many effects on the oceans and the marine ecosystems for decades or centuries to come. These challenges will co-exist with present pressures from human activities, such as eutrophication, shipping, fishing and use of coastal areas for harbours, tourism and recreation. The proposed project is designed to aid developments of protection and management plans by providing essential information on climate-change impacts on one of the most polluted, regional seas in the World: The Baltic Sea. We will extend an existing database of climate-impact scenarios by the use of a regional climate model for downscaling of global climate models and IPCC global climate change scenarios. In climate change research, ensembles of climate simulations are produced in an attempt to cover the uncertainty in future projections. However, due to limited computer resources, regional studies can most often not make use of the full ensemble and important information on the spread between models will be lost. We will here exploit the extended database in order to develop a method that will enable an optimum subset ensemble that will capture the uncertainty in the projections, while minimizing the number of models in the downscaling. The method will be performed with respect to user requirements of climate-change impact studies and for key marine climate and environmental variables. It will be done within a framework general enough to be transferred to other regional seas.