Repository logo

Hierarchical Bayesian analysis and statistical learning theory II: water management application




Khalil, Abedalrazq, author
McKee, Mac, author
U.S. Committee on Irrigation and Drainage, publisher

Journal Title

Journal ISSN

Volume Title


Water scarcity and uncertainties in forecasting future water availabilities present serious problems for basin-scale water management. These problems create a need to design intelligent prediction models that learn and adapt to their environment in order to provide water managers with decision-relevant information related to the operation of river systems. State-of-the-art techniques fused into a model paradigm (described in Part I of this manuscript) will be demonstrated as decision tools to enhance real-time water management. The framework previously discussed in Part I will be able to diagnose abnormality in the system. Abnormality in this context is referred to as outliers, false signals (e.g., the result of sensor failure) and system behavior "drift" (i.e., non-stationarity or "concept drift"). The proposed versatile adaptive paradigm might be utilized in any control process of a dynamical system in which a quantitative characterization of uncertainty is required. The utility and practicality of this proposed approach is demonstrated here with an application in a real case study river basin.


Presented during the USCID water management conference held on October 13-16, 2004 in Salt Lake City, Utah. The theme of the conference was "Water rights and related water supply issues."

Rights Access



Associated Publications