Global Climate Models (GCMs) can provide essential meteorological data as inputs for simulating and assessing the impact of climate change on catchment hydrology. However, downscaling of GCM outputs is often required due to their coarse spatial and temporal resolution. As an effective downscaling method, stochastic weather generators can reproduce daily sequences with statistically similar statistical characteristics. Most weather generators can only simulate single-site meteorological data, which are spatially uncorrelated. Therefore, this study introduces a method for multi-site precipitation downscaling based on a combination of a single-site stochastic weather generator, CLIGEN (CLImate GENerator), and a modified shuffle procedure constrained with multi-model ensemble GCM monthly precipitation outputs. The applicability of the downscaling method is demonstrated in the Huangfuchuan Basin (arid to semi-arid climate) for a historical period (1976–2005) and a projection period (2021–2070, historical, the representative concentration path (RCP) 2.6, RCP4.5, RCP4.8 scenarios) to generate spatially correlated daily precipitation. The results show that the proposed downscaling method can accurately simulate the mean of daily, monthly and annual precipitation and the wet spell lengths, and the inter-station correlation among 10 sites in the basin. In addition, this combination method generated the projected precipitation and showed an increasing trend for future years. These findings could help us better cope with the potential risks of climate change.

中文翻译：

---

使用基于 GCM 的月度数据进行每日降水生成的多站点统计降尺度方法

全球气候模式 (GCM) 可以提供必要的气象数据，作为模拟和评估气候变化对流域水文影响的输入。然而，由于空间和时间分辨率粗糙，通常需要对 GCM 输出进行缩小。作为一种有效的降尺度方法，随机天气发生器可以重现具有统计相似统计特征的日常序列。大多数天气发生器只能模拟单站点气象数据，这些数据在空间上是不相关的。因此，本研究介绍了一种基于单站点随机天气发生器 CLIGEN (CLImate GENerator) 和受多模式集合 GCM 月降水输出约束的改进洗牌程序的组合的多站点降水降尺度方法。在皇甫川盆地（干旱到半干旱气候）的历史时期（1976-2005）和预测期（2021-2070，历史，代表性浓度路径（RCP）2.6，RCP4）证明了降尺度方法的适用性.5、RCP4.8 情景）以生成空间相关的日降水量。结果表明，所提出的降尺度方法可以准确模拟流域内10个站点的日、月和年降水量平均值和湿期长度以及站间相关性。此外，这种组合方法产生了预测的降水量，并在未来几年呈现出增加的趋势。这些发现可以帮助我们更好地应对气候变化的潜在风险。