【摘要】 基于黄河上游有关水文气象台站的降水径流资料,将LocalModeling方法应用于龙羊峡水库月入库径流预报的中长期水文预报模型.模型的检验和应用结果表明,该方法有着稳健性好、数学物理意义明确、对数据系列要求不高和容易操作等优点,在非汛期各月的径流预测中具有较高的准确性,并且在考虑了降水的影响后,对汛期径流的计算精度亦基本符合水文情报预报规范和实际应用的要求.该模型在黄河上游水量预报和调度工作中具有良好的应用前景.更多还原

【Abstract】 The Longyangxia waterpower plant is the first large waterpower project in the upper Yellow River. Its reservoir possesses a capacity of 274×10⁸ km³ and 95% of the inflow to the reservoir mainly come from the drainage above Tangnag located in the northeastern Tibetan Plateau between 95.5°～103’5°E and 32’5°～36’0°N, with a water collection area of 12’19×10⁴ km², accounting for 1/6 of the total area of the Yellow River Basin. However, the runoff in the upper Yellow River has been decreasing due to global warming and human activity since the end of the 1990s, which not only great affects the economy and people living in the upper Yellow River areas, but also restrains the economic development of whole Yellow River Basin. So accurate predicting the future runoff in the upper Yellow River is indispensable to adequate and reasonable exploitation of the water resources in the basin, as well as determining the amount of water supplied from other basin. In this paper, based on runoff series at the Tangnag Hydrometric Station and Local Modeling method, one of n-dimension and non-linear dynamical system forecast method, a monthly runoff-forecasting model is presented to predict the monthly inflow to the reservoir. Practice shows that the Local Modeling method possesses some marked advantages such as possessing favorable stability, possessing clear mathematics and physics meanings, being prone to operating and so on, as compared with some deterministic and non- deterministic forecast method. The model fits especially for forecasting monthly runoff in low-flow period and forecasting monthly runoff in flood season with satisfied veracity, if the effect of precipitation is calculated. So the model is thought to accord with the requirement of hydrology forecast criterion and becomes an available forecast.更多还原