Projected changes in climatic extremes, compared to the mean climate, exhibit a greater negative impact on the natural environment. Several studies reported that multi-model ensemble approach can improve the reliability of hydro-climatic extreme projection by extracting important information from a large number of general circulation models (GCMs). However, most of the available multi-model assembling methods do not consider both the spatial and temporal variabilities. Thus, this study reflects both the spatial and temporal climate characteristics during multi-model averaging through the Taylor diagram skill metrics. The capability of the proposed multi-model assembling approach was evaluated for reproducing the multitude of climate extreme indices. Moreover, the reliability of a multi-model assembling approach was assessed for preserving the maximum variability of the GCMs output. In general, the results showed that multi-model assembling approach outperformed the individual climate models for reproducing the hydro-climatic extremes; however, it artificially corrupted and narrowed the projected climate extremes variability of the GCMs output. Thus, it is worthwhile to consider both the individual climate models and multi-model ensemble projections toward an improved projection of hydro-climatic extremes. In general, the study proved that the impacts of climate change on the hydro-climatic extremes are more amplified compared to the changes in mean climate. Hence, this study suggests that meaningful efforts should be put in the future to proactively manage the risks of climate extremes.

与平均气候相比，预计的极端气候变化将对自然环境产生更大的负面影响。几项研究报告说，多模型集成方法可以通过从大量的常规循环模型（GCM）中提取重要信息来提高水文气候极端投影的可靠性。但是，大多数可用的多模型组装方法都没有同时考虑空间和时间变化。因此，本研究通过泰勒图技能指标反映了多模型平均过程中的时空气候特征。对所提出的多模型组合方法的能力进行了评估，以再现多种极端气候指数。此外，为了保留GCM输出的最大可变性，评估了多模型组装方法的可靠性。总体而言，结果表明，在再现水文气候极端事件方面，多模型组合方法的性能优于单个气候模型。但是，它人为地破坏和缩小了GCM产出的预计极端气候变化性。因此，有必要同时考虑单个气候模型和多模型集合投影，以改善水文气候极端情况。总体而言，该研究证明，与平均气候变化相比，气候变化对极端水文气候的影响更大。因此，这项研究表明，未来应该做出有意义的努力，以主动管理极端气候的风险。