本文基于黄河流域1979-2018年的ERA-Interim再分析气象与水文数据，以及CMIP5中10个气候模式下3种典型浓度路径（RCP2.6、RCP4.5、RCP8.5）的全球气候变化数据，采用离散的数据处理方法，建立了黄河流域贝叶斯网络模型，推断了黄河流域近40余年来气候要素对径流的影响概率，预测了黄河流域未来径流量。结果表明:1979-2018年黄河天然径流量呈减小趋势，基于贝叶斯网络分区间概率预测预报的径流量也呈减小趋势；黄河流域的不同区间（高、中、低）径流量对气候的敏感程度不同，但径流始终与降水相关性最高；在RCP2.6情景下，黄河流域未来20年、60年的径流量为585.50亿m3，588.57亿m3；在RCP4.5情景下，其值为585.42亿m3，587.53亿m3；在RCP8.5情景下，其值为593.50亿m3，585.11亿m3。

This paper is based on the reanalyzed meteorological and hydrological data of the Yellow River Basin from 1979 to 2018, and the global climate data of 3 typical concentration paths (RCP2.6, RCP4.5, RCP8.5) under 10 climate models in CMIP5, using discrete Data processing methods are used to establish a Bayesian network model of the Yellow River Basin, infer the probability of the impact of climate factors on the runoff of the Yellow River Basin in the past 40 years, and predict the future runoff of the Yellow River Basin. The results showed that the natural runoff of the Yellow River from 1979 to 2018 showed a decreasing trend, and the predicted runoff based on Bayesian network inter-zone probability also showed a decreasing trend; the Yellow River basin differed Interval (high, medium, and low) runoff has different sensitivity to climate, but runoff is always the most correlated with precipitation; Under the RCP2.6 scenario, the Yellow River Basin’s runoff in the next 20 and 60 years will be 58.55 billion m3, 58.857 billion m3, under the RCP4.5 scenario the value is 58.542 billion m3, 58.753 billion m3, and under the RCP8.5 scenario the value is 59.35 billion m3 , 58.511 billion m3.

国家重点研发计划项目(2017YFC0404303);