洪水预报模型的预评估机模型研究

洪水过程是一个高度复杂的过程，目前还没有一种水文模型是最优的，本文首次提出了洪水预报模型预评估机的概念，对给定的洪水预报模型的预报效果进行预评估，以作为是否采信该模型预报结果和选择洪水预报模型的依据之一。本文建立了基于相似系数和神经网络的预评估机模型，并通过对实例的分析，对它们实际应用的可行性进行了探讨，其中基于相似系数的预评估机模型尚有待改进；基于神经网络的预评估机模型有良好的应用前景。

Discussion on pre-evaluation machine of flood forecast models

The concept of pre-evaluation machine(PEM)of flood forecast model is introduced.By applying PEM,the performance of certain flood forecast model using certain input can be evaluated,so whether the result is reliable or not can be decided,and the output of PEM can be a reference to model selecting as well.Two types of PEM model is presented,the feasibility and reliablity of them are discussed.