Evaluation of modeled changes in extreme precipitation in Europe and the Rhine basin

Abstract

In this study, we investigate the change in multi-day precipitation extremes in late winter in Europe using observations and climate models. The objectives of the analysis are to determine whether climate models can accurately reproduce observed trends and, if not, to find the causes of the difference in trends.Similarly to an earlier finding for mean precipitation trends, and despite a lower signal to noise ratio, climate models fail to reproduce the increase in extremes in much of northern Europe: the model simulations do not cover the observed trend in large parts of this area. A dipole in the sea-level pressure trend over continental Europe causes positive trends in extremes in northern Europe and negative trends in the Iberian Peninsula. Climate models have a much weaker pressure trend dipole and as a result a much weaker (extreme) precipitation response.The inability of climate models to correctly simulate observed changes in atmospheric circulation is also primarily responsible for the underestimation of trends in the Rhine basin. When it has been adjusted for the circulation trend mismatch, the observed trend is well within the spread of the climate model simulations. Therefore, it is important that we improve our understanding of circulation changes, in particular related to the cause of the apparent mismatch between observed and modeled circulation trends over the past century.

A statistical framework for conditional extreme event attribution

“…On the one hand, one looks at a persisting zonal circulation in a rather broad sense. On the other hand, one looks at a circulation trajectory that looks like the observation of January 2014, which yielded an atypical zonal pattern (van Oldenborgh et al, 2015). …”

“…A probability of sufficient causation is defined by 1 − 1−p 1 1−p 0 . An alternative approach to factual/counterfactual worlds can be proposed, as in van Haren et al (2013): a "new" world in which we live, like the recent decades, and an "old" world in which our ancestors lived, like the beginning of the 20th century. We implicitly assume that these two worlds are different (at least from the environmental point of view).…”

Local impact analysis of climate change on precipitation extremes: are high-resolution climate models needed for realistic simulations?

“…However, these results need to be validated against historical precipitation observations prior to any use for local impact studies of climate change. When GCM results are validated based on observations, sometimes large biases are observed, especially for extreme precipitation values (van Pelt et al, 2012;van Haren et al, 2013;Tabari et al, 2015), imposing an uncertainty on the GCM projections for the future. The biases in the coarse-resolution GCMs come from the fact that they disregard some governing features of precipitation at local scale, next to the scale differences when comparing GCM results with local observations (Maraun et al, 2010;Willems et al, 2012).…”

Incorporating circulation statistics in bias correction of GCM ensembles: hydrological application for the Rhine basin

“…The CP-method leads to a larger improvement in simulated discharge in the Alpine area in Precipitation bias in climate models is primarily related to coarse resolution, lack of ability to simulate explicitly local processes, misrepresentation of physical processes, all of which can in some areas be amplified by feedbacks between climate components (Wang et al 2014). In Central Europe, model biases in sea surface temperature contribute to the precipitation bias, although precipitation changes in this area are primarily caused by circulation changes (van Ulden and van Oldenborgh 2006;van Haren et al 2013). Depending on the area of interest and the physical processes dominating the precipitation biases, we set up a framework to perform a correction of the circulation dependent precipitation bias.…”