Short-term water level forecasting in the Yangtze River is of significant importance for the strategies regarding flood prevention, inland shipping and domestic water supply. An ARIMA (Autoregressive Integrated Moving Average) model is thus proposed for daily water level forecasting at three stations in the middle reach of the Yangtze River. The model is trained and calibrated using a historic dataset of four years (2012–2015). Predictive performance of the ARIMA model is evaluated statistically by a dataset of year 2016. Satisfactory model predictions have been obtained in the study area although some discrepancies have been observed when sharp fluctuations of water level occur. The model accuracy is explored in terms of root mean square error (RMSE), mean absolute percentage error (MAPE), percent of bias (PBIAS) and the index of agreement (d). Moreover, the influence of forecasting period on the model accuracy is further investigated. It is found that the model accuracy decreases as the forecasting period is extended. Considering the simplicity and efficiency of the ARIMA, it provides an effective way of short term water level forecasting and shows potential to be extended for alternative research aims.
ARIMA modelling and forecasting of water level in the middle reach of the Yangtze River
2017-08-01
720820 byte
Conference paper
Electronic Resource
English
PMs concentration forecasting using ARIMA algorithm
IEEE | 2020
|HENRY – Federal Waterways Engineering and Research Institute (BAW) | 2014
|HENRY – Federal Waterways Engineering and Research Institute (BAW) | 2008
|