The Development of a Hybrid Wavelet-ARIMA-LSTM Model for Precipitation Amounts and Drought Analysis

Investigation of quantitative predictions precipitation amounts and forecasts drought events are conducive to facilitating early warnings. However, there has been limited research into or modern statistical analyses over Northeast China, one the most important grain production regions. Therefore, a case study at three meteorological sites which represent different climate types was explored, we used time series analysis monthly grey theory methods for annual during 1967–2017. Wavelet transformation (WT), autoregressive integrated moving average (ARIMA) long short-term memory (LSTM) were utilized depict series, new hybrid model wavelet-ARIMA-LSTM (W-AL) developed. In addition, GM (1, 1) DGM China Z-Index (CZI) based on introduced forecast events, because system specializes in small sample results poor information. The revealed that (1) W-AL exhibited higher prediction accuracy forecasting than ARIMA LSTM; (2) CZI values calculated through suggested more slight occurred Changchun while moderate frequently Linjiang Qian Gorlos; (3) performed better event forecasting.