Probabilistic forecasts of near-term climate change based on a resampling ensemble technique

Probabilistic forecasts of near-term climate change are derived by using a multimodel ensemble of climate change simulations and a simple resampling technique that increases the number of realizations for the possible combination of anthropogenic climate change and internal climate variability. The technique is based on the assumption that the probability distribution of local climate changes is only a function of the all-model mean global average warming. Although this is unlikely to be exactly true, cross-verification indicates that the resulting biases are more than compensated by the increased sample size provided by the method. The resulting forecasts for southern Finland suggest a 95% probability of annual mean warming and an 80% probability of increased annual mean precipitation from the years 1971–2000 to 2011–2020 under the SRES A1B emissions scenario. The choice of the emissions scenario is unimportant for such short-term forecasts but becomes gradually more important towards the late 21st century. The simulations also suggest that the probability of near-term warming that is large enough to be discernible from internal variability is largest in the tropics where internal temperature variability is small, not in the Arctic where the average model-simulated warming is largest.