人工神经网络用于水文资料的插补延长

The application of artificial neural networks to the extrapolation of hydrological data

水文资料的插补延长一直是水文计算中的一个难题 .采用人工神经网络建立两个水文相似流域之间的耦合模型 ,用一个流域的资料来推求另一个流域的径流量 .另外人工神经网络还用来捕捉降雨与径流间潜在的关系 ,从另一途径解决水文资料的插补延长问题 .大量的数值实验表明 ,人工神经网络可以成功地用于水文资料的外插或无资料地区的径流模拟预测 .

The extension of hydrological series is one of the difficult problems which hydrologists always encounter. This paper will adopt artificial neural networks (ANNs) to set up a hybrid model of two hydrological similar catchments. The data from a hydrological similar catchment is used to train an ANN model in order to calculate the runoff of another catchment. In addition, the relation between rainfall and the increment of runoff is also established to provide an alternative way to solve the extrapolation problem. Amount of numerical experiments show that ANNs can be applied successfully to extend the hydrological data or to perform runoff simulation of catchment without hydrological data.