en Large-scale climate oscillations have been known to have a direct impact into the variability of hydrological systems and their relationship have been documented in different parts of the world. Those indices play an important role in the hydrological time series behaviour, thus identifying their relationship is an important tool for an accurate time series forecast, especially under uncertain hydrological time series and multiobjective demands from reservoirs. This paper proposes to use random forests, a powerful ensemble learning technique, to evaluate the relative relationship of each climate index to the streamflow’s time series. Such nonlinear machine learning methods have been successfully used in different areas of knowledge. Then, a regression decision tree is built aiming to extract useful information about dry and wet periods from streamflow data related to the climate oscillation and adjust the decision tree model to forecast future values of streamflow. The data used in this study belongs to the gauge station of Itaipu’s hydropower plant located in the Paraná River (Brazil). The basin has an annual cycle of discharges with large-amplitude and high flows in the summer.
%0 Journal Article
%1 rolim2019exploring
%A Rolim, LZR
%A Filho, FA Souza
%A Rocha, RV
%A Reis, GA
%A Carvalho, TMN
%D 2019
%K imported topic_earthenvironment
%T Exploring the relationship between climate indices and hydrological time series using a machine learning approach
%X en Large-scale climate oscillations have been known to have a direct impact into the variability of hydrological systems and their relationship have been documented in different parts of the world. Those indices play an important role in the hydrological time series behaviour, thus identifying their relationship is an important tool for an accurate time series forecast, especially under uncertain hydrological time series and multiobjective demands from reservoirs. This paper proposes to use random forests, a powerful ensemble learning technique, to evaluate the relative relationship of each climate index to the streamflow’s time series. Such nonlinear machine learning methods have been successfully used in different areas of knowledge. Then, a regression decision tree is built aiming to extract useful information about dry and wet periods from streamflow data related to the climate oscillation and adjust the decision tree model to forecast future values of streamflow. The data used in this study belongs to the gauge station of Itaipu’s hydropower plant located in the Paraná River (Brazil). The basin has an annual cycle of discharges with large-amplitude and high flows in the summer.
@article{rolim2019exploring,
abstract = {[en] Large-scale climate oscillations have been known to have a direct impact into the variability of hydrological systems and their relationship have been documented in different parts of the world. Those indices play an important role in the hydrological time series behaviour, thus identifying their relationship is an important tool for an accurate time series forecast, especially under uncertain hydrological time series and multiobjective demands from reservoirs. This paper proposes to use random forests, a powerful ensemble learning technique, to evaluate the relative relationship of each climate index to the streamflow’s time series. Such nonlinear machine learning methods have been successfully used in different areas of knowledge. Then, a regression decision tree is built aiming to extract useful information about dry and wet periods from streamflow data related to the climate oscillation and adjust the decision tree model to forecast future values of streamflow. The data used in this study belongs to the gauge station of Itaipu’s hydropower plant located in the Paraná River (Brazil). The basin has an annual cycle of discharges with large-amplitude and high flows in the summer.},
added-at = {2024-11-29T11:56:28.000+0100},
author = {Rolim, LZR and Filho, FA Souza and Rocha, RV and Reis, GA and Carvalho, TMN},
biburl = {https://puma.scadsai.uni-leipzig.de/bibtex/2c7d6098bdb15ce7937918f402e6c26f9/joum576e},
citation = {11th World Congress on Water Resources and Environment: Managing Water …, 2019},
interhash = {65c8e86a405bbb4f9250af2a351cd220},
intrahash = {c7d6098bdb15ce7937918f402e6c26f9},
keywords = {imported topic_earthenvironment},
timestamp = {2024-11-29T11:56:28.000+0100},
title = {Exploring the relationship between climate indices and hydrological time series using a machine learning approach},
year = 2019
}
