大幅度打破记录的气候事件对社会和生态系统构成威胁。预计气候变化会增加此类事件发生的可能性,但是由于自然的可变性和有限的数据可用性,量化这些概率是具有挑战性的,特别是对于观察和非常罕见的极端。在这里,我们估算了至少一个工业前标准偏差的范围内破坏记录的降水事件的概率。使用大型合奏气候模拟和极值理论,我们确定了经验和分析记录破碎的概率,发现它们非常吻合。我们证明,特别是在高排放情况下,模型预测,到本世纪末,几乎所有全球土地在相对于静止气候的变化中,创纪录的降水概率要高得多,热带地区的脆弱地区增长最强劲。我们证明,不断增加的变异性是近期创纪录的降水概率增加的重要驱动因素,并提出了一个框架,该框架可以量化均值和变异性的综合趋势对极端降水中破纪录行为的影响。在变暖的世界中,创纪录的降水事件的概率估计对于为风险评估和适应政策提供信息至关重要。
Climate events that break records by large margins are a threat to society and ecosystems. Climate change is expected to increase the probability of such events, but quantifying these probabilities is challenging due to natural variability and limited data availability, especially for observations and very rare extremes. Here we estimate the probability of precipitation events that shatter records by a margin of at least one pre-industrial standard deviation. Using large ensemble climate simulations and extreme value theory, we determine empirical and analytical record shattering probabilities and find they are in high agreement. We show that, particularly in high emission scenarios, models project much higher record-shattering precipitation probabilities in a changing relative to a stationary climate by the end of the century for almost all the global land, with the strongest increases in vulnerable regions in the tropics. We demonstrate that increasing variability is an essential driver of near-term increases in record-shattering precipitation probability, and present a framework that quantifies the influence of combined trends in mean and variability on record-shattering behaviour in extreme precipitation. Probability estimates of record-shattering precipitation events in a warming world are crucial to inform risk assessment and adaptation policies.