Application of an improved time serie smodel to flow and sediment forecasting in a non-flood period
Lian Ji-jian · Water Resources and Hydropower Engineering · 2004
The improvement of the conventional method of error real-time correction based on ARMA model and the prolongation of the forecasting range of time series using normal distribution instate of purely random series are presented in this article. These two methods were used at Tongguan hydrometric station to test flow and sediment forecasting and the results computed have shown efficiently improved accuracy of prediction.