Local-scale changes in mean and heavy precipitation in Western Europe, climate change or internal variability?

• Published in: Climate dynamics 2018-06, Vol.50 (11-12), p.4745-4766
• Main Authors: Aalbers, Emma E., Lenderink, Geert, van Meijgaard, Erik, van den Hurk, Bart J. J. M.
• Format: Article
• Language: English
• Subjects: Atmospheric models; Chaos theory; Climate change; Climate models; Climate system; Climate variability; Climatology; Computer simulation; Daily precipitation; Earth and Environmental Science; Earth Sciences; Environmental aspects; Environmental impact; Extreme weather; Geophysics/Geodesy; Global warming; Greenhouse effect; Heavy precipitation; High resolution; Noise; Oceanography; Precipitation; Precipitation variability; Probability theory; Regional analysis; Regional climate models; Regional climates; Resolution; Variability
• ISSN: 0930-7575; 1432-0894
• DOI: 10.1007/s00382-017-3901-9

High-resolution climate information provided by e.g. regional climate models (RCMs) is valuable for exploring the changing weather under global warming, and assessing the local impact of climate change. While there is generally more confidence in the representativeness of simulated processes at higher resolutions, internal variability of the climate system—‘noise’, intrinsic to the chaotic nature of atmospheric and oceanic processes—is larger at smaller spatial scales as well, limiting the predictability of the climate signal. To quantify the internal variability and robustly estimate the climate signal, large initial-condition ensembles of climate simulations conducted with a single model provide essential information. We analyze a regional downscaling of a 16-member initial-condition ensemble over western Europe and the Alps at 0.11° resolution, similar to the highest resolution EURO-CORDEX simulations. We examine the strength of the forced climate response (signal) in mean and extreme daily precipitation with respect to noise due to internal variability, and find robust small-scale geographical features in the forced response, indicating regional differences in changes in the probability of events. However, individual ensemble members provide only limited information on the forced climate response, even for high levels of global warming. Although the results are based on a single RCM–GCM chain, we believe that they have general value in providing insight in the fraction of the uncertainty in high-resolution climate information that is irreducible, and can assist in the correct interpretation of fine-scale information in multi-model ensembles in terms of a forced response and noise due to internal variability.