Premium
Streamflow Forecasting Using Singular Value Decomposition and Support Vector Machine for the Upper Rio Grande River Basin
Author(s) -
Bhandari Swastik,
Thakur Balbhadra,
Kalra Ajay,
Miller William P.,
Lakshmi Venkat,
Pathak Pratik
Publication year - 2019
Publication title -
jawra journal of the american water resources association
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.957
H-Index - 105
eISSN - 1752-1688
pISSN - 1093-474X
DOI - 10.1111/1752-1688.12733
Subject(s) - streamflow , geopotential height , climatology , environmental science , forecast skill , lead time , structural basin , drainage basin , meteorology , precipitation , geology , geography , cartography , paleontology , marketing , business
The current study improves streamflow forecast lead‐time by coupling climate information in a data‐driven modeling framework. The spatial–temporal correlation between streamflow and oceanic–atmospheric variability represented by sea surface temperature (SST), 500‐mbar geopotential height (Z 500 ), 500‐mbar specific humidity (SH 500 ), and 500‐mbar east–west wind (U 500 ) of the Pacific and the Atlantic Ocean is obtained through singular value decomposition (SVD). SVD significant regions are weighted using a nonparametric method and utilized as input in a support vector machine (SVM) framework. The Upper Rio Grande River Basin (URGRB) is selected to test the applicability of the proposed model for the period of 1965–2014. The April–August streamflow volume is forecasted using previous year climate variability, creating a lagged relationship of 1–13 months. SVD results showed the streamflow variability was better explained by SST and U 500 as compared to Z 500 and SH 500 . The SVM model showed satisfactory forecasting ability with best results achieved using a one‐month lead to forecast the following four‐month period. Overall, the SVM results showed excellent predictive ability with average correlation coefficient of 0.89 and Nash–Sutcliffe efficiency of 0.79. This study contributes toward identifying new SVD significant regions and improving streamflow forecast lead‐time of the URGRB.