中文摘要：

为评估水文模型预测气候变化影响下水文过程演变的可靠性,以淮河紫罗山流域为研究对象,利用1977-1989年的逐日观测资料,基于三水源新安江模型分析了不同率定期下模型参数值和模拟精度的变化以及与降水量变化特征的关系.结果表明：不同率定期参数率定值具有较大的不确定性,但影响蒸散发量、产流量、水源划分以及地下水出流速率的参数值组合与降水系列的干湿变化存在明显的相关关系;以确定性系数和均方根误差作为目标函数评估模拟效果时发现,两者都随着验证期日降水量均方差的增大而增大,同时更偏向于提高降水变幅大的丰水年组的预测效果.

英文摘要：

In order to evaluate hydrological process simulated by hydrological models under climate impact,rainfall-runoff process of Zilu-oshan catchment in Huaihe basin was selected and analyzed with daily observation data. In this paper,calibrated model parameters and simu-lation accuracy were analyzed under different calibration periods between 1977 and 1989,revealing their relationships with precipitation vari-ation characteristics. The results show that the calibrated parameters change dramatically among different calibration periods. However,there are obvious correlations between precipitation series and parameter combinations affecting evapotranspiration,runoff,division of runoff com-ponents and groundwater runoff rate;when choosing deterministic coefficient and root mean square error as objective functions to evaluate the simulation results,they both increase with the growth of square mean error of daily precipitation,and parameters calibrated from wet periods with high mean square error of precipitation series result in more steady simulation accuracy.