近日，华南理工大学王兆礼团队报道了极端复合事件对累积CO2排放的增强响应。该项研究成果发表在2026年5月13日出版的《自然》杂志上。
复合事件（例如同时发生的湿热与干热极端事件）是地球上最具影响的气候灾害之一，预计在全球变暖背景下将变得更加严重。尽管瞬时平均温度对累积CO2排放的响应已得到良好量化，但复合事件的相应响应仍不甚明确。
研究组表明，复合事件对累积CO2排放的瞬时响应（定义为每单位累积CO2排放所引发的事件频率变化）强烈依赖于复合事件的背景发生频率。具体而言，研究组发现，历史上频繁发生的复合事件随累积CO2排放的增加几乎呈线性增长，而较为罕见且更严重的事件则以不成比例的速度加剧。此外，经观测约束的TCoRE比多模型平均值高出37%–75%，表明复合极端事件的发生频率将高于地球系统模型的预测。
该约束还将模型集合的不确定性降低了37%–56%。应用约束后的TCoRE进一步表明，在考虑复合事件变化的情况下，与将升温限制在1.5°C和2°C目标相一致的允许CO2排放量将显著降低。研究组提出TCoRE作为一个简单、稳健且受观测约束的指标，与气候风险评估和政策制定具有直接相关性。
附：英文原文
Title: Enhanced response of extreme compound events to cumulative CO2 emissions
Author: Li, Jun, Zhang, Yao, Ciais, Philippe, Zhang, Hongying, Wang, Zhaoli, Tang, Hongwu, Piao, Shilong
Issue&Volume: 2026-05-13
Abstract: Compound events—such as concurrent hot–wet and drought–heat extremes—are among the most consequential climate hazards on Earth1,2,3,4 and are projected to become more severe under warming. Although the transient mean temperature response to cumulative CO2 emissions has been well quantified5,6,7,8, the corresponding response of compound events remains less clear. Here we show that the response of the transient compound events to cumulative CO2 emissions (TCoRE), defined as the change in event frequency per unit of cumulative CO2 emissions, is strongly dependent on the background frequency of compound events. In particular, we find that historically frequent compound events increase almost linearly with increasing cumulative CO2 emissions, whereas rarer and more severe events escalate disproportionately. Moreover, the observationally constrained TCoRE is 37–75% higher than the multi-model average, indicating that compound extremes will occur more frequently than Earth system models project. The constraint also reduces model ensemble uncertainty by 37–56%. Applying the constrained TCoRE further suggests that the allowable CO2 emissions consistent with limiting warming to 1.5°C and 2°C are substantially lower when accounting for changes in compound events. We propose the TCoRE as a simple, robust and observationally constrained metric with direct relevance for climate risk assessment and policy development.
DOI: 10.1038/s41586-026-10544-1