Although uncertainties are still large, many potentially dangerous effects have already been identified concerning the impacts of global warming on human societies. For example, the record-breaking 2003 summer heat wave in Europe has given a glimpse of possible future European climate conditions. Here we use an ensemble of regional climate simulations for the end of the twentieth and twenty-first centuries over Europe to show that frequency, length and intensity changes in warm and cold temperature extremes can be derived to a close approximation from the knowledge of changes in three central statistics, the mean, standard deviation and skewness of the Probability Distribution Function, for which current climate models are better suited. In particular, the effect of the skewness parameter appears to be crucial, especially in the case of cold extremes, since it mostly explains the relative warming of these events compared to the whole distribution. An application of this finding is that the future impacts of extreme heat waves and cold spells on non-climatological variables (e.g., mortality) can be estimated to a first-order approximation from observed time series of daily temperature transformed in order to account for simulated changes in these three statistics.
Heat Wave Probability Distribution Function Cold Spell Temperature Time Series Extreme Tail
These keywords were added by machine and not by the authors. This process is experimental and the keywords may be updated as the learning algorithm improves.
This is a preview of subscription content, log in to check access.
Supplementary Figure 1Ratio changes (scenario/control) in multi-model mean length of tail events for detrended daily TMEAN. A warm (cold) tail event is defined as a set of consecutive warm (cold) tail days. Results are shown for M-A2 (blue), MS-A2 (green), MSW-A2 (red) and A2 (black). Ratio changes are first averaged for each grid-point in Europe and then for each model. Discontinuous colored areas display a measure of model uncertainty, defined here as one inter-model standard deviation above and below the multi-model mean. (JPEG 1.44 MB).