基于径流-蒸散发-蓄水量协同率定的平原区水文模拟研究

Research on Hydrological Simulation in Plain Areas Based on Collaborative Calibration of Runoff,Evapotranspiration and Water Storage

平原区无统一流域出口断面,相关模型参数率定与水文模拟是水文学研究的薄弱环节和难点问题。以黄淮海平原为研究对象,基于实测数据、遥感反演结果及同化的区域水循环通量和储量等多源要素数据集,利用可变下渗容量(VIC)模型,构建了基于径流、蒸散发和蓄水量协同率定的模型参数率定方法,评估了其在黄淮海平原水文模拟中的适用性,并利用机器学习、广义似然不确定性估计等方法分析了该模型对水文模拟精度、模型参数异参同效性和径流模拟不确定性等方面的改进效果。得到的主要结论为:①VIC模型在黄淮海平原具有较好的适用性,与实测值相比,模拟径流和蒸散发量在率定期的相关系数大于0.9,蓄水量的相关系数大于0.6;②对于VIC模型的6个参数,第2层土壤厚度d 2最敏感,其次是可变下渗曲线指数B,其余参数敏感性较弱;③与传统的单一径流率定方法相比,参数协同率定方法提高了蒸散发量和蓄水量的模拟精度,减小了模型参数的优选空间与不确定性置信区间,模拟径流的90%置信区间约缩小26%,降低了水文模拟的不确定性。研究成果对水文模拟技术进步具有重要的支撑作用。

There is no unified watershed outlet section in the plain area,and its model parameter calibration and hydrological simulation are the weak link and difficult problem in hydrology research.Taking the Huang-Huai-Hai Plain as the research object,based on measured data,remote sensing inversion results,and assimilated multi-source factor datasets of regional water cycle flux and reserves,a variable infiltration capacity(VIC)model was used to construct a model parameter calibration method based on collaborative calibration of runoff,evapotranspiration,and water storage.Its applicability in hydrological simulation of the Huang-Huai-Hai Plain was evaluated.With methods such as machine learning and genera-lized likelihood uncertainty estimation,the improvement effect of the model on the accuracy of hydrological simulation,the same effect of different parameters of model parameters and the uncertainty of runoff simulation were analyzed.The research results are as follows.Firstly,the VIC model has good applicability in the Huang-Huai-Hai Plain,with a correlation coefficient of over 0.9 between simulated runoff and evapotranspiration compared to measured values,and a correlation coefficient of over 0.6 for water storage.Secondly,for the six parameters of the VIC model,the second layer soil thickness d 2 is the most sensitive,followed by the variable infiltration curve index B,and the sensitivity of the other parameters is relatively weak.Thirdly,compared with the traditional single runoff calibration method,the parameter calibration method of collaborative calibration improves the simulation accuracy of evapotranspiration and water storage,and reduces the optimization space and uncertainty confidence interval of model parameters.The 90%confidence interval of simulated runoff has been reduced by approximately 26%,reducing the uncertainty of hydrological simulation.The research results play an important supporting role in the progress of hydrological simulation technology.