Climate models predict that East Asia (EA) will be substantially warmer than the present despite large inter-model uncertainty. This study investigated the major sources of the climate projections and the inter-model uncertainty. Particularly, we decomposed the differences in surface temperatures between the historical and RCP8.5 runs from 26 CMIP5 into partial surface temperature changes due to individual radiative and non-radiative processes through the climate feedback-response analysis method. Results show that anthropogenic greenhouse forcing and subsequent water vapor feedback processes are primarily responsible for the surface warming over EA. Relatively more rapid warming over the snow/ice-covered area and southern China is due to feedback processes associated with surface albedo and cloud, respectively. The regional warming is, however, compensated by the surface non-radiative (sensible and latent heat) cooling. The inter-model projection uncertainty is substantially large over high latitudes and the Tibetan Plateau mainly due to surface albedo feedback. Again, this large uncertainty is partly suppressed by surface non-radiative cooling. Water vapor and cloud feedbacks are the secondary important sources of the projection uncertainty. Moreover, the contributions of greenhouse forcing and atmospheric dynamics to the projection uncertainty are found to be minor.

气候模型预测，尽管模型间存在较大不确定性，但东亚（EA）将会比目前的温​​度更高。这项研究调查了气候预测和模型间不确定性的主要来源。特别是，通过气候反馈响应分析方法，我们将历史数据与RCP8.5运行数据之间的地表温度差异（从26 CMIP5）分解为部分表面温度变化（由于各个辐射和非辐射过程）。结果表明，人为温室强迫和随后的水蒸气反馈过程是造成EA地表变暖的主要原因。在冰雪覆盖地区和华南地区，相对较快的升温是由于分别与地表反照率和云有关的反馈过程。但是，区域性变暖是 通过表面非辐射（显热和潜热）冷却进行补偿。在高纬度地区和青藏高原上，模型间投影的不确定性很大，这主要是由于地表反照率的反馈。同样，这种较大的不确定性部分通过表面非辐射冷却得到了抑制。水蒸气和云的反馈是预测不确定性的次要重要来源。此外，发现温室强迫和大气动力学对预测不确定性的贡献很小。水蒸气和云的反馈是预测不确定性的次要重要来源。此外，发现温室强迫和大气动力学对预测不确定性的贡献很小。水蒸气和云的反馈是预测不确定性的次要重要来源。此外，发现温室强迫和大气动力学对预测不确定性的贡献很小。