Dynamic runoff simulation in a changing environment: A data stream approach

被引：28

作者：

Yang, Qinli ^{[1
,2
,3
]}

Zhang, Heng ^{[1
]}

Wang, Guoqing ^{[3
]}

Luo, Shasha ^{[1
]}

Chen, Dongzi ^{[1
]}

Peng, Wanshan ^{[1
]}

Shao, Junming ^{[2
,4
]}

机构：

[1] Univ Elect Sci & Technol China, Sch Resources & Environm, 2006 Xiyuan Ave, Chengdu 611731, Sichuan, Peoples R China

[2] Univ Elect Sci & Technol China, Big Data Res Ctr, 2006 Xiyuan Ave, Chengdu 611731, Sichuan, Peoples R China

[3] Nanjing Hydraul Res Inst, State Key Lab Hydrol Water Resources & Hydraul En, Nanjing 210029, Jiangsu, Peoples R China

[4] Univ Elect Sci & Technol China, Sch Comp Sci & Engn, 2006 Xiyuan Ave, Chengdu 611731, Sichuan, Peoples R China

来源：

ENVIRONMENTAL MODELLING & SOFTWARE | 2019年 / 112卷

基金：

中国国家自然科学基金;

关键词：

Runoff; Climate change; Land cover change; Data stream mining; CLIMATE-CHANGE; SACRAMENTO MODEL; SURFACE RUNOFF; WATER; SOIL; PARAMETERS; CATCHMENT; HYDROLOGY; SELECTION; COVER;

D O I：

10.1016/j.envsoft.2018.11.007

中图分类号：

TP39 [计算机的应用];

学科分类号：

081203 ; 0835 ;

摘要：

In this study, we introduce a data stream method for dynamic runoff simulation, which allows capturing the evolving relationship between runoff and its impact factors (e.g., temperature, rainfall). The basic idea is to view continuously arriving data of runoff and its impact factors as a data stream, and dynamically learn its relationship. To validate the effectiveness of the proposed method, we compare its performance with that of three data driven models (ANN, SVR, Random Forest) and six representative hydrological models (SWAT, AWBM, SimHyd, SMAR, Sacramento, and Tank) in simulating monthly runoff. The proposed method performs well with the best NSE of 0.88, being superior to comparable models. Furthermore, the data stream model also shows its advantage in the flexibility of combing various impact factors of runoff into the model. The findings demonstrate that the data stream method provides a promising way to dynamically simulate runoff in a changing environment.

引用

页码：157 / 165

页数：9

共 67 条

[1] NEW LOOK AT STATISTICAL-MODEL IDENTIFICATION [J].

AKAIKE, H .

IEEE TRANSACTIONS ON AUTOMATIC CONTROL, 1974, AC19 (06) :716-723

[2]

Allen R. G., 1998, FAO Irrigation and Drainage Paper

[3]

Almeida Ezilda, 2013, Machine Learning and Knowledge Discovery in Databases. European Conference, ECML PKDD 2013. Proceedings: LNCS 8188, P480, DOI 10.1007/978-3-642-40988-2_31

[4] Large area hydrologic modeling and assessment - Part 1: Model development [J].