Climate in Sweden during the past millennium – Evidence from proxy data, instrumental data and model simulations

This report concerns a study which was conducted for SKB. The conclusions and viewpoints presented in the report are those of the authors and do not necessarily coincide with those of the client. Summary Knowledge about climatic variations is essential for SKB in its safety assessments of a geological repository for spent nuclear waste. There is therefore a need for information about possible future climatic variations under a range of possible climatic states. However, predictions of future climate in any deterministic sense are still beyond our reach. We can, nevertheless, try to estimate the magnitude of future climate variability and change due to natural forcing factors, by means of inferences drawn from natural climate variability in the past. Indeed, the climate of the future will be shaped by the sum of natural and anthropogenic climate forcing, as well as the internal climate variability. The aim here is to review and analyse the knowledge about Swedish climate variability, essentially during the past millennium. Available climate proxy data and long instrumental records provide empirical information on past climatic changes. We also demonstrate how climate modelling can be used to extend such knowledge. We use output from a global climate model driven with reconstructed radiative forcings (solar, volcanic and greenhouse gas forcing), to provide boundary conditions for a regional climate model. The regional model provides more details of the climate than the global model, and we develop a simulated climate history for Sweden that is complete in time and space and physically consistent. We use output from a regional model simulation for long periods in the last millennium, to study annual mean temperature , precipitation and runoff for the northern and southern parts of Sweden. The simulated data are used to place corresponding instrumental records for the 20 th century into a plausible historical perspective. We also use output from the regional model to study how the frequency distribution of the daily temperature, precipitation, runoff and evaporation at Forsmark and Oskarshamn could have varied between unusually warm and cold 0-year periods during the last millennium. Models, however, cannot be used to deduce the exact time evolution of climate variations, but they can provide relevant information in a statistical sense, for example by defining the limits within which climate naturally has varied. Uncertainties – and also advantages – of both empirical climate data and model simulations are discussed. A main conclusion is that there …