珠江三角洲网河区顶点分水分沙变化及神经网络模型预测

作者根据西、北江干流高要、石角及西北江三角洲网河区顶点马口、三水共4个水文站的流量、输沙率的月均序列进行统计，探讨近几十年来西北江三角洲网河区顶点分水分沙的季节变化和多年变化。分析结果表明，网河区顶点的分水分沙格局以1993年为转折点，自1993年后，水沙分配已发生了重大变化，三水站的分流比和分沙比突然增大，分流比约增大了一倍，保持在20％以上，分沙比在16％以上。应用非线性神经网络模型对马口、三水两水文站的分流比和分沙比进行建模，并对网河区顶点的分水分沙变化趋势作多步预测，结果表明，1998～2000年三水站的分流比均在 25％以上；1998年的分沙比较小，预测值为 12％，1999年和 2000年的预测值分别是15％、16％。

The change of diversion ratio of flow and sediment in the Pearl River Delta and application of the neural network prediction model

The authors collected hydrologic data at 4 stations(Gaoyao and Shijiao in the main stream of the West River and the North River and Makou and Sanshui at the head of network in the Peal River Delta).The seasonal and annual changes of diversion ratios of flow discharge and sediment transport were analyzed.The results indi cated that the 2 diversion ratios had changed significantly since 1993.The two diversion ratios at Sanshui station increased abruptly and they were greater than 20% and 16% respectively.An artificial neural network model was applied to predict the diversion ratios of flow and sediment at Makou and Sanshui station.The predicted diversion ratio of flow was greater than 25% during the period 1998 to 2000.The predicted diversion ratio of sediment was 12% in 1998,15% in 1999 and 16% in 2000.