Research on a coupling model for groundwater depth forecasting
Chao Song, Xianqi Zhang, Dengkui Hu, Wei Tuo · Desalination and Water Treatment · 2019
ABSTRACT Groundwater depth forecasting plays an important role in agricultural irrigation, rational utilization of soil and water resources and ecological protection. The groundwater resource system is influenced by multiple factors such as temperature, precipitation, evapotranspiration, surface water recharge and groundwater discharge, it is characterized by randomness and non-stationary. The empirical mode decomposition (EMD) can decompose the signal into sub-signals of different frequencies and can reduce the non-stationary of the original signal, the Elman neural network (ENN) has strong nonlinear approximation ability. Based on the characteristics of the above two methods, the EMD- Elman coupling forecasting model was constructed and applied to groundwater depth forecasting in People’s Victory Canal Irrigation District. The results show that EMD-Elman model has better forecasting effect and lower forecasting error and is better than single back propagation neural network model and ENN model. Furthermore, under human over-exploitation, the forecasting accuracy of the EMD-Elman model will be slightly reduced, but the forecasting effect is still in the acceptable range. This research has an important guiding value on revealing People’s Victory Canal Irrigation District of groundwater dynamic change rule and provide a new way for groundwater depth forecasting.