基于土壤属性的VIC模型基流参数估计框架
Estimation of baseflow parameters in variable infiltration capacity model using soil properties
-
摘要: "异参同效"现象是水文模型参数率定过程中的一个难题。为了减少需要率定的模型参数个数,从而降低模型参数的相互作用,构建了利用土壤属性直接估计VIC(Variable Infiltration Capacity)模型3个基流参数的框架,并在3个位于不同水文气候条件下的典型流域中作了实例研究。传统的VIC模型参数估计方法需要率定6个参数;而在新的参数估计框架下,需要率定的模型参数从6个减少为3个。蒙特卡罗模拟结果表明新的参数估计框架提高了参数的敏感性。同时,新的参数估计方法模拟的径流过程和之前通过率定得到的径流过程差别很小。GLUE(Generalised Likelihood Uncertainty Estimation)方法分析得到径流模拟的不确定性结果表明:新的参数估计框架计算径流的90%置信区间要明显小于传统方法,而且对于低流量的模拟效果更加明显,也就是说采用新的参数估计框架可以显著降低径流模拟的不确定性。这种基流参数估计方法可以被应用于其他相似的水文模型中。Abstract: Equifinality is a difficult but unavoidable challenge in hydrologic modeling.A framework is established and presented to estimate the three baseflow parameters in the Variable Infiltration Capacity (VIC) model using soil properties.The new parameter estimation method is tested for streamflow simulations on three distinct catchments in hydro-climatic conditions in China.The simulations are compared to the previous calibrated results.The new method can reduce the number of VIC calibration parameters from six to three, thus decreasing model equifinality and uncertainty.The sensitivity of the calibrated parameters under the new framework is improved as demonstrated by the Monte Carlo simulations.The parameter uncertainty can also be reduced.In addition, the uncertainty of VIC simulations is assessed using the Generalised Likelihood Uncertainty Estimation (GLUE) method.Results show that the uncertainty of streamflow simulations (or, confidence interval) is low with the new parameter estimation method compared to the uncertainty resulted from the previous calibration method.The new method is applicable to the VIC model and other similar hydrologic models for streamflow simulations.