Local Modeling模式及其在月径流预测中的应用

Local Modeling方法是一种动力系统预测方法，将其应用于河西内陆区黑河干流出山口莺落峡水文站月平均流量的中长期预测预报，取得了较为理想的成果。预测试验的结果表明，该预测模型有较高计算精度，尤其适用于非主汛期各月的月平均流量的预测；对于主汛期6～9月的月平均流量的预测，在考虑前期来水与预见期内降水的影响后，亦可获到较为理想的预测结果。可以认为，该方法的预报精度达到了水文情报预报规范的要求，Local Modeling方法的应用，将为西北干旱地区河川径流的中长期预报提供了一个新的途径。

Local Modeling Method and Its Application in Monthly Runoff Prediction in the Northwest Arid Inland Region

The occurrence and development of hydrology phenomena are the result of many factors interacting together and among them climate is a main influencing factor. On condition that precipitation and air temperature data in the mountain area of the northwestern China were seldom recorded, the runoff status at the mountain outlet could reflect the changes of climate environment in the mountain area. The basic idea of Local Modeling method is that rivers are regarded as multi-dimensional nonlinear dynamical system controlled by hydrology, climate and physical geography etc, in which runoff is the relatively important factor. It was considered in model that the future status of the system consist in the historical data, from the historical data the spots most similar to the current status of the system can be found. That is, the model finds the most adjacent spots to the current status vector in status vector spaces determined by historical data in time sequence; then using the obtained distances and the next status of adjacent spots to estimate the future status of the system. Proved by application in the forecast for the monthly runoff of Heihe river the method possesses a good forecast precision,especially is applicable for the runoff forecast in non-main flood seasons. As for the runoff forecast in main flood seasons, a satisfying forecast precision still can be obtained in premise of considering the influence of precipitation in the forecast period and runoff in last month. So it can be thought that the forecasting precise based on Local Modeling method has met the demands of Standard for Hydrological Information and Hydrological Forecasting announced by the National Water Conservancy Ministry, and the method will provide a new path for river runoff forecast of the arid region in northwest China.