Long‐term streamflow forecasting using artificial neural network based on preprocessing technique

• DOI: http://doi.org/10.1002/for.2564
• Published date: 01 April 2019
• Author: Zhi‐Yu Wang,Fang‐Fang Li,Jun Qiu
• Date: 01 April 2019

RESEARCH ARTICLE

Long‐term streamflow forecasting using artificial neural

network based on preprocessing technique

Fang‐Fang Li

1

| Zhi‐Yu Wang

2

| Jun Qiu

3

1

College of Water Resources and Civil

Engineering, China Agricultural

University, Beijing 100083, China

2

Shandong Water Conservancy

Vocational College, Rizhao 276826, China

3

State Key Laboratory of Hydroscience

and Engineering, Tsinghua University,

Beijing 100084, China

Correspondence

Jun Qiu, State Key Laboratory of

Hydroscience and Engineering, Tsinghua

University, Beijing 100084, China.

Email: aeroengine@tsinghua.edu.cn

Funding information

National Key R&D Program of China,

Grant/Award Number: 2017YFC0403600,

2017YFC0403602; Open Project of State

Key Laboratory of Plateau Ecology and

Agriculture, Qinghai University, Grant/

Award Number: 2016‐KF‐03

Abstract

Artificial neural network (ANN) combined with signal decomposing methods is

effective for long‐term streamflow time series forecasting. ANN is a kind of

machine learning method utilized widely for streamflow time series, and which

performs well in forecasting nonstationary time series without the need of phys-

ical analysis for complex and dynamic hydrological processes. Most studies take

multiple factors determining the streamflow as inputs such as rainfall. In this

study, a long‐term streamflow forecasting model depending only on the histori-

cal streamflow data is proposed. Various preprocessing techniques, including

empirical mode decomposition (EMD), ensemble empirical mode decomposi-

tion (EEMD) and discrete wavelet transform (DWT), are first used to decompose

the streamflow time series into simple components with different timescale

characteristics, and the relation between these components and the original

streamflow at the next time step is analyzed by ANN. Hybrid models EMD‐

ANN, EEMD‐ANN and DWT‐ANN are developed in this study for long‐term

daily streamflow forecasting, and performance measures root mean square error

(RMSE), mean absolute percentage error (MAPE) and Nash–Sutcliffe efficiency

(NSE) indicate that the proposed EEMD‐ANN method performs better than

EMD‐ANN and DWT‐ANN models, especially in high flow forecasting.

KEYWORDS

artificial neural network (ANN), discrete wavelet translate (DWT),empirical mode decomposition

(EMD), ensemble empirical mode decomposition (EEMD), hybridmodel, long‐term streamflow

forecasting

1|INTRODUCTION

Long‐term streamflow forecasting plays an important role

in water resources management and utilization (Yang,

Tian, Sun, Yuan, & Hu, 2014; Yaseen et al., 2016), such

as the assessment of water consumption, water supply,

drought and flood disasters, and the scheduling of water

resources.

A number of research studies have been reported on

hydrological forecasting in past decades, mainly including

physical models and mathematical methods (X. L. Zhang,

Peng, Zhang, & Wang, 2015). Physical models explore the

hydrological dynamic process of watershed combing

weather processes (Chen & Brissette, 2015; Smiatek,

Kunstmann, & Werhahn, 2012; Ye et al., 2017),

meteorological conditions (Hanna et al., 2013; Ralph,

Coleman, Neiman, Zamora, & Dettinger, 2013) and under-

lying surface conditions (Rosenberg, Clark, Steinemann, &

Lettenmaier, 2013; Sinha, Sankarasubramanian, &

Mazrooei, 2014). However, due to the complexity of the

hydrological processes affected by various factors such as

runoff, rainfall and human activity, not only a large

Received: 2 June 2018 Revised: 26 September 2018 Accepted: 21 November 2018

DOI: 10.1002/for.2564

192 © 2018 John Wiley & Sons, Ltd. Journal of Forecasting. 2019;38:192–206.wileyonlinelibrary.com/journal/for