Influence of internal variability on future changes in extreme wind speed over North America

Internal variability plays an important role in the projection of climate and extreme climate changes at regional scales. Here, we use large ensemble experiments from two different climate models [the Canadian Earth System Model Large Ensemble (CanESM2-LE) with 50 members and the Community Earth System Model Large Ensemble (CESM1-LE) with 40 members to investigate the influence of internal variability on the projection of extreme wind speed (extreme WS) in the future (2081–2100) relative to the present climatology (1986–2005) over North America. Generally, the spatial patterns of future changes in extreme WS across members in both ensembles exhibit quite large diversity for the whole year and winter, while consistent changes among members are found in summer. This indicates that the influence of internal variability on the future changes in extreme WS is higher in winter than summer. With enough members, the forced responses allow direct comparisons to demonstrate the model uncertainty. Increases in wintertime extreme WS at high latitudes and around the Hudson Bay, as well as decreases over the west-central USA and along the southwest coast of the domain, are largely influenced by the external forcing with small model uncertainty. Regions within the 40°–60°N latitudinal band of the domain are projected to experience external forced decrease in extreme WS in summer with high model consistency. Results from high-resolution of dynamical downscaling large ensemble simulations (CanRCM4-LE) are comparable with those of its driving models (CanESM2-LE) and can provide detailed information in regions with sea–land contrast and complex topography, which is of great vital for the local policy maker.