Multivariate Bayesian Regression Approach to Forecast Releases from a System of Multiple Reservoirs

Abstract

This research presents a model that simultaneously forecasts required water releases 1 and 2 days ahead from two reservoirs that are in series. In practice, multiple reservoir system operation is a difficult process that involves many decisions for real-time water resources management. The operator of the reservoirs has to release water from more than one reservoir taking into consideration different water requirements (irrigation, environmental issues, hydropower, recreation, etc.) in a timely manner. A model that forecasts the required real-time releases in advance from a multiple reservoir system could be an important tool to allow the operator of the reservoir system to make better-informed decisions for releases needed downstream. The model is developed in the form of a multivariate relevance vector machine (MVRVM) that is based on a sparse Bayesian regression model approach. With this Bayesian approach, a predictive confidence interval is obtained from the model that captures the uncertainty of both the model and the data. The model is applied to the multiple reservoir system located in the Lower Sevier River Basin near Delta, Utah. The results show that the model learns the input-output patterns with high accuracy. Computing multiple-time-ahead predictions in real-time would require a model which guarantees not only good prediction accuracy but also robustness with respect to future changes in the nature of the inputs data. A bootstrap analysis is used to guarantee good generalization ability and robustness of the MVRVM. Test results demonstrate good performance of predictions and statistics that indicate robust model generalization abilities. The MVRVM is compared in terms of performance and robustness with another multiple output model such as Artificial Neural Network (ANN). A. M. Ticlavilca (B) · M. McKee

Particle Swarm Optimization for Automatic Calibration of Large Scale Water Quality Model (CE-QUAL-W2): Application to Karkheh Reservoir, Iran

“…As the system complexities increases, the computational and memory requirements for large number of function evaluation may restrict its performance and form a computational burden. Coupling the proposed methodology in this research with machine learning techniques (Cheng et al 2002;Muttil and Chau 2006;Ticlavilca and McKee 2010;Nazif et al 2010) to estimate system responses may form a suitable approach to break the computational bottleneck.…”

Optimizing embedded sensor network design for catchment-scale snow-depth estimation using LiDAR and machine learning

“…This process would be well‐suited to the field of machine learning, which includes “unsupervised” algorithms (which identify patterns in the independent variables without observing the dependent variable), and “supervised” or “active” algorithms (which employ limited observations of the independent variable in order to improve the estimator). A number of recent studies have applied machine‐learning algorithms to topics in hydrology, such as runoff and streamflow estimation [ Solomatine and Shrestha , ; Londhe and Charhate , ], evapotranspiration modeling [ Torres et al ., ], streamflow forecasting [ Rasouli et al ., ], assessment of the contamination of groundwater [ Khalil et al ., ], and estimation of needs for reservoir releases [ Khalil et al ., ; Ticlavilca and McKee , ].…”

Wavelet-multivariate relevance vector machine hybrid model for forecasting daily evapotranspiration

“…single output. When predictions are to be made for multiple steps in time, multiple outputs, the multivariate relevance vector machine (MVRVM) reported by Thayananthan et al (2008) has proven more effective (Ticlavilca and McKee 2011;Torres et al 2011;Ticlavilca et al 2013).…”