ABSTRACT Rainfall events largely control hydrological processes occurring on and in the ground, but the performance of climate models in reproducing rainfall events has not been investigated enough to guide selection among the models when making hydrological projections. We proposed to compare the durations, intensities, and pause periods, as well as depths of rainfall events when assessing the accuracy of general circulation models (GCMs) in reproducing the hydrological characteristics of observed rainfall. We also compared the sizes of design storm events and the frequency and severity of drought to demonstrate the consequences of GCM selection. The results show that rainfall and extreme hydrological event projections could significantly vary depending on climate model selection and weather stations, suggesting the need for a careful and comprehensive evaluation of GCM in the hydrological analysis of climate change. The proposed methods are expected to help to improve the accuracy of future hydrological projections for water resources planning.

中文翻译：

---

评估气候模型在再现降雨事件的水文特征方面的性能

摘要 降雨事件在很大程度上控制着地面上和地面上发生的水文过程，但气候模型在再现降雨事件方面的性能尚未得到足够的研究，以指导在进行水文预测时在模型之间进行选择。我们建议在评估一般环流模型 (GCM) 再现观测降雨水文特征的准确性时，比较降雨事件的持续时间、强度和停顿期以及深度。我们还比较了设计风暴事件的规模以及干旱的频率和严重程度，以证明 GCM 选择的后果。结果表明，降雨和极端水文事件预测可能会因气候模型选择和气象站的不同而有显着差异，表明需要在气候变化的水文分析中对 GCM 进行仔细和全面的评估。预计所提出的方法将有助于提高未来水资源规划水文预测的准确性。