“…Precipitation projections suffer from large uncertainties partly due to the poor representation of dynamical mechanisms and the difficulties to correctly parametrize the internal variability of the climate system in GCMs(van Haren, Oldenborgh, Lenderink, Collins, & Hazeleger, 2013a;van Haren, Oldenborgh, Lenderink, & Hazeleger, 2013b). Furthermore, variations in patterns such as NAO, poorly represented in the simulated climate data as shown inFigure 6f,j, are linked to other Earth subsystems such as the North Atlantic Jet (NAJ;Trouet, Babst, & Meko, 2018); the Sahara/Sahel and West African Monsoon (WAM;Bader & Latif, 2003); and the El Niño-Southern Oscillation (ENSO;Kalimeris, Ranieri, Founda, & Norrant, 2017).Those systems may shift with increased anthropogenic forcing(Barnes & Polvani, 2013) and pass critical states in intensity and location, leading to large-scale impacts on human and ecological systems (socalled tipping elements;Lenton et al, 2008; see Schwalm et al, 2017 for a worldwide study). Thus, changes in NAJ, WAN and ENSO may promote profound changes in the state of the NAO and, hence, in the precipitation over the Mediterranean region.Although the poor representation of dynamical mechanisms inGCMs is still a relevant issue, recent studies relying on more accurate representations of fine-scale circulation patterns have shown improved predictive skills(van Haren, Haarsma, Oldenborgh, & Hazeleger, 2015;Ozturk et al, 2015).…”