Ensembles of regional-global climate model combinations show substantial spread in projected wind and solar resources. Using 31 RCM-GCM pairs, we quantify the sources of this spread with a spatially and seasonally resolved variance decomposition, separating contributions from RCMs and GCMs. For both wind speed and solar radiation, RCMs dominate the variability in the absolute historical fields. In contrast, projected changes in wind speed are largely controlled by the driving GCMs, except in mountainous regions where RCM-induced variance becomes larger than that induced by GCMs. For solar radiation, contributions are strongly season-dependent, with RCMs dominating in summer and GCMs in winter. Our findings support that GCM and RCM variability together define the uncertainty of wind and solar climate projections. This provides guidance for designing climate model ensembles that better support uncertainty-aware energy system decisions under climate change.

翻译：区域-全球气候模型组合的集合在预估风能与太阳能资源时展现出显著离散性。基于31组区域气候模型-全球气候模型配对，我们通过空间与季节分解的方差量化方法厘清了该离散性的来源，分离了区域气候模型与全球气候模型的贡献。对于风速与太阳辐射而言，区域气候模型在绝对历史场中的变异性中占主导地位。相比之下，风速的未来变化预估主要受驱动全球气候模型控制，但在山区区域，区域气候模型引发的方差超过全球气候模型。对于太阳辐射，贡献呈现强季节性依赖——夏季以区域气候模型主导，冬季以全球气候模型为主。我们的研究表明，全球气候模型与区域气候模型的变异性共同定义了风能与太阳能气候预估的不确定性，这为设计能更好支撑气候变化下不确定性感知型能源系统决策的气候模型集合提供了指导。