Detection of anthropogenic influence on the evolution of record-breaking temperatures over Europe

Abstract

Changes in temperature extreme events are expected as a result of anthropogenic climate change, but uncertainties exist in when and how these changes will be manifest regionally. This is especially the case over Europe due to different methodologies and definitions of temperature extreme events. An alternative approach is to examine changes in record-breaking temperatures. Datasets of observed temperature combined with ensembles of climate model simulations are used to assess the possible causes and significance of record-breaking temperature changes over the late twentieth and twenty-first centuries. A simple detection methodology is first applied to evaluate the extent to which the effect of anthropogenic forcing can be detected in present-day observed and simulated changes in record-breaking temperature. We then study the projected evolution of record-breaking daily minimum and maximum temperatures over the twenty-first century in Europe with a climate model. The same detection approach is used to identify the time of emergence of the anthropogenic signal relative to a model-derived estimate of internal variability. From the 1980s onwards, a change in the evolution of cold and warm records is observed and simulated, but it still remains in the range of internal variability until the end of the twentieth century. Minimum and maximum record-breaking temperatures tend to occur (respectively) less and more often than during the 1960s and 1970s taken as representative of a stationary climate. Model simulations with natural forcing only fail to reproduce the observed changes after the 1980s while the latter are compatible with simulations constrained by anthropogenic forcings. The deviation from the characteristic behavior of a stationary climate record-wise initiated in the 1980s is projected to accentuate during the twenty-first century. Annual changes become inconsistent with the model-derived internal variability between the 2020s and 2030s. Over the last three decades of the twenty-first century and under the RCP8.5 scenario, warm records occur on average five times more often than initially. Conversely, breaking new cold record become extremely difficult. The Mediterranean region is particularly affected in summer, whereas central and northeastern Europe is more impacted in winter.