Observational constraint methods based on the relationship between the past global warming trend and projected warming across climate models were used to reduce uncertainties in projected warming by the Intergovernmental Panel on Climate Change. Internal climate variability in the eastern tropical Pacific associated with the so-called pattern effect weakens this relationship and has reduced the observed warming trend over recent decades. Here we show that regressing out this variability before applying the observed global mean warming trend as a constraint results in higher and narrower twenty-first century warming ranges than other methods. Whereas the Intergovernmental Panel on Climate Change assessed that warming is unlikely to exceed 2 °C under a low-emissions scenario, our results indicate that warming is likely to exceed 2 °C under the same scenario, and hence, limiting global warming to well below 2 °C will be harder than previously anticipated. However, the reduced uncertainties in these projections could benefit adaptation planning.

政府间气候变化专门委员会使用基于过去全球变暖趋势与气候模型预测变暖之间关系的观测约束方法来减少预测变暖的不确定性。热带东部太平洋的内部气候变化与所谓的模式效应相关，削弱了这种关系，并减少了近几十年来观测到的变暖趋势。在这里，我们表明，在应用观察到的全球平均变暖趋势作为约束之前，回归这种变异性会导致比其他方法更高和更窄的二十一世纪变暖范围。尽管政府间气候变化专门委员会评估在低排放情景下升温不太可能超过 2°C，但我们的结果表明在相同情景下升温可能超过 2°C，因此将全球变暖限制在远低于2°C 将比之前预期的更加困难。然而，这些预测中不确定性的减少可能有利于适应规划。