对流域水文模型的再认识

讨论了流域水文模型的本质,指出模型只是原型的近似,不可能反映原型的一切特征,尤其像流域降雨径流形成这样复杂的现象。诠释了"分解"与"集成"的科学思维在流域水文模型建模中的作用。认为只有产汇流理论和集成基本水文现象为流域水文模型的方法有所突破,才会有流域水文模型的实质性进步。分析了流域水文模型反问题不适定的原因,提出了直接最优化方法用于解决流域水文模型反问题时应注意的事项,以及用"大数据"方法解决流域水文模型反问题的设想和初步思路。认为随着水文观测手段的进一步精准,"大数据"方法在解决流域水文模型反问题和正问题中均有美好的应用前景。

More discussion of watershed hydrological model

The nature of watershed hydrological model is discussed. Models are always approximations for prototypes and a model cannot reflect all the characteristics of a prototype, especially for the complicated phenomena such as the formation of rainfall-runoff of a watershed. The role of scientific thinking for decomposition and assembly in the establishment of watershed hydrological models is explained. Substantial progress of watershed hydrological models can be made only if the breakthrough of both the theories for runoff yield and concentration and the assembly methods from basic hydrological phenomenon to watershed hydrological models are realized. The reasons of ill-posed problems in the inverse problems of watershed hydrological models are analyzed. Some shortcomings of the solution for inverse problems of watershed hydrological models by means of direct optimization methods are pointed out. Preliminary ideas on the solution for the inverse problems of watershed hydrological models through big data are proposed. With more accurate and refined hydrological observation, big data method has bright prospects in the solution for both direct and inverse problems of watershed hydrological models.