Quantitative assessment of natural internal variability and externally forced responses of Northern Hemisphere (NH) temperatures is necessary for understanding and attributing climate change signals during past warm and cold periods. However, it remains a challenge to distinguish the robust internally generated variability from the observed variability. Here, large-ensemble (70 member) simulations, Energy Balance Model simulation, temperature ensemble reconstruction, and three dominant external forcings (volcanic, solar, and greenhouse gas) were combined to estimate the internal variability of NH summer (June–August) temperatures over the past 2000 years (1–2000 CE). Results indicate that the Medieval Climate Anomaly was predominantly attributed to centennial-scale internal oscillation, accounting for an estimated 104% of the warming anomaly. In contrast, the Current Warm Period is influenced mainly by external forcing, contributing up to 90% of the warming anomaly. Internal temperature variability offsets cooling by volcanic eruptions during the Late Antique Little Ice Age. These findings have important implications for the attribution of past climate variability and improvement of future climate projections.

对北半球（NH）温度的自然内部变率和外部强迫响应进行定量评估对于理解和归因过去温暖和寒冷时期的气候变化信号是必要的。然而，区分强大的内部产生的变异性和观察到的变异性仍然是一个挑战。在此，将大型系综（70 名成员）模拟、能量平衡模型模拟、温度系综重建和三种主要外部强迫（火山、太阳和温室气体）相结合，以估计过去 2000 年（公元 1-2000 年）北半球夏季（6 月至 8 月）温度的内部变化。结果表明，中世纪气候异常主要归因于百年尺度的内部振荡，估计占变暖异常的 104%。相比之下，当前暖期主要受外强迫影响，对变暖异常的贡献率高达90%。内部温度变化抵消了小冰河时代晚期火山喷发带来的降温。这些发现对于过去气候变化的归因和未来气候预测的改进具有重要意义。