This study aimed to investigate the influence of hydrological model calibration/validation on discharge projections for three large river basins (the Rhine, Upper Mississippi and Upper Yellow). Three hydrological models (HMs), which have been firstly calibrated against the monthly discharge at the outlet of each basin (simple calibration), were re-calibrated against the daily discharge at the outlet and intermediate gauges under contrast climate conditions simultaneously (enhanced calibration). In addition, the models were validated in terms of hydrological indicators of interest (median, low and high flows) as well as actual evapotranspiration in the historical period. The models calibrated using both calibration methods were then driven by the same bias corrected climate projections from five global circulation models (GCMs) under four Representative Concentration Pathway scenarios (RCPs). The hydrological changes of the indicators were represented by the ensemble median, ensemble mean and ensemble weighted means of all combinations of HMs and GCMs under each RCP. The results showed moderate (5–10%) to strong influence (> 10%) of the calibration methods on the ensemble medians/means for the Mississippi, minor to moderate (up to 10%) influence for the Yellow and minor (< 5%) influence for the Rhine. In addition, the enhanced calibration/validation method reduced the shares of uncertainty related to HMs for three indicators in all basins when the strict weighting method was used. It also showed that the successful enhanced calibration had the potential to reduce the uncertainty of hydrological projections, especially when the HM uncertainty was significant after the simple calibration.

中文翻译：

---

气候变化下水文模型校准对预测水文变化的影响——三大流域的多模型评估

本研究旨在调查水文模型校准/验证对三个大型流域（莱茵河、密西西比河上游和黄河上游）的流量预测的影响。三个水文模型（HMs），首先根据每个流域出口的月流量进行校准（简单校准），同时在对比气候条件下根据出口和中间仪表的日流量重新校准（增强校准） . 此外，这些模型在感兴趣的水文指标（中值、低流量和高流量）以及历史时期的实际蒸散量方面得到了验证。然后，使用两种校准方法校准的模型由来自四个代表性浓度路径情景 (RCP) 下的五个全球环流模型 (GCM) 的相同偏差校正气候预测驱动。指标的水文变化以每个RCP下HMs和GCMs所有组合的集合中值、集合均值和集合加权均值表示。结果显示校准方法对密西西比州的集合中位数/均值有中度 (5–10%) 到强烈影响 (> 10%)，对黄色和次要 (< 5 %) 对莱茵河的影响。此外，当使用严格加权方法时，增强的校准/验证方法减少了所有流域三个指标与 HMs 相关的不确定性份额。