Too‐stable North Atlantic climate system in CMIP6 experiments undermines precipitation projections in Europe

Abstract This study investigates fluctuations in the spatial pattern of the North Atlantic Oscillation (NAO) during the Industrial Era. NAO variability was analyzed over successive 30‐year periods, clustering its spatial patterns using two metrics: the gradient ( g ) between its action centres and flux direction (fd). Three distinct NAO patterns were identified, each exhibiting significant implications for NAO–precipitation modulation in the North Atlantic region. Coupled Model Intercomparison Project (CMIP6) climate models were evaluated for their ability to replicate these observed patterns. While many models accurately reproduce all three patterns, they frequently misinterpret their frequencies, favouring a disproportionately stable NAO. This bias, consistent across historical, natural‐only, and greenhouse gas forcings (GHG only), as well as piControl simulations, appears intrinsic to the models. This misinterpretation is significantly reduced in ssp585 scenarios, suggesting a potential link between forcing levels and improved multidecadal NAO variability in models. The occurrence of those different NAO patterns introduces uncertainty in precipitation projections, as this bias can persist in future scenarios. These findings highlight the need for improved model representation of NAO spatial variability to enhance the reliability of precipitation projections, particularly for Europe.