Hybrid Models Combining EMD/EEMD and ARIMA for Long-Term Streamflow Forecasting

Abstract

Long-term streamflow forecast is of great significance for water resource application and management. However, accurate monthly streamflow forecasting is challenging due to its non-stationarity and uncertainty. Time series analysis methods have been proved to perform well in stationary time series forecasting, which can be derived from decomposition of the non-stationary sequence. As common decomposition methods in time domain, Empirical Mode Decomposition (EMD) and Ensemble Empirical Mode Decomposition (EEMD) are selected to decompose the components with different time-scale characteristics in the original hydrological time series in this study. The derived components are proved to be stationary by the stationarity test. Thus, Autoregressive Integrated Moving Average (ARIMA) model, a simple and effective time series analysis method, is used to forecast the components. A hybrid EMD/EEMD-ARIMA model is proposed in this study for long-term streamflow forecasting, which is applied to the upper stream of the Yellow River. The original daily streamflow time series of six years at the Tangnaihai station are firstly decomposed by EMD/EEMD into several stationary or simple non-stationary sub-series to explore detailed data information with different time scales. ARIMA models with appropriate parameters are then established for each subsequence to forecast the stream flow of the next year. Predicted ten-day and monthly stream flow is finally obtained combing the predictions of all the components. The EMD-ARIMA hybrid model performs best in forecasting high and moderate value of streamflow and fits best with the observation compared with EEMD-ARIMA and ARIMA models. The results not only verify the effectiveness of the proposed hybrid EMD/EEMD-ARIMA model in exploiting comprehensive information to improve the prediction but also indicate that the EMD-ARIMA model with end points disposal performs the best and can be used for long-term hydrological forecasting.

Hybrid Deep Learning Predictor for Smart Agriculture Sensing Based on Empirical Mode Decomposition and Gated Recurrent Unit Group Model

“…Liu et al [64] implemented a hybrid method to predict wind speed, in which the wavelet transform was used to decompose the wind speed signal into two components and the approximated signal (one of two components) was modeled by a support vector machine. Zhi et al [65] selected empirical mode decomposition (EMD) as the decomposition method for the time series, and components with different features in the original hydrological time series were decomposed. Yaslan et al [66] predicted power load demand using a combined EMD and support vector regression (SVR) model.…”

Development of Multidecomposition Hybrid Model for Hydrological Time Series Analysis

“…The objective of EMD is to decompose the nonlinear and nonstationary data adaptively into number of oscillatory components called Intrinsic Mode Decomposition (IMF). A number of studies have been conducted combining the EMD with data driven models [15,[18][19][20][21]. Specifically in hydrology, EMD is used with ANN for wind speed and stream flow prediction [15,20].…”

Peer Review #2 of "Improving the prediction accuracy of river inflow using two data pre-processing techniques coupled with data-driven model (v0.1)"

“…Accurate prediction results in better decisions such as, flood and drought controls, the supply of drinking water, water resources management and physical-based models, data-driven models, and hybrid models (Chen et al, 2018). All these models have been widely used to predict rivers flow and other hydrologic analyses (Erdal & Karakurt, 2013;Hao et al, 2017;Chen et al, 2018;Darwen, 2018;Wang et al, 2018). Physicalbased models extract the inherent behaviors of hydrological variables by conceptualizing their physical process and characteristics.…”

“…They took monthly streamflow data and concluded that application of ARIMA can be useful in generating precise prediction. However, the disadvantages of TS models are that the river inflow data must be linear which limits the application of these models (Wang et al, 2018). (Behzad et al, 2009) and many other non-linear problems (Wu & Lin, 2019).…”

“…However, all data pre-processing techniques have some drawbacks with different aspects to decompose the non-linear, multi-scale and noisiest data. The most widely used WA depends heavily on the selection of wavelet basis function (Wang et al, 2018), the application of EMD is limited by its own mathematical mode mixing and sensitivity to denoise property (Nazir et al, 2019), and EEMD suffers a strict mathematical theory (Qian et al, 2019;Wu et al, 2019).…”