Document Type
Presentation
Publication Date
8-1-2014
Abstract
A methodology is developed for reservoir release decisions considering forecasted downstream dissolved oxygen local conditions. River water quality management using reservoirs focuses mainly on how to develop a release schedule that may improve downstream conditions based on the seasonal change of the water quality within the reservoir. This improvement, however does not take into account the downstream local water quality state, which in certain cases might be more important, as the pollutant load downstream could be diluted with the upstream available volume released from the reservoir. Field sampling collected data suggest that the dissolved oxygen concentration decay produced by polluted tributary inflows to the main stream, may be reduced . Development of long term operation guide curves combined with short- to medium-range water quality forecasts for the downstream conditions was tested as a possible solution. In order to include the water quality state in the reservoir release decision, a forecasting meta-model was developed using M5 model trees algorithm. This data-driven model is able to recreate the state of the system while reducing the number of input variables of the numerical model. After testing the methodology it was concluded that the main stream water quality can actually be improved with the long term operation rules based on downstream water quality conditions. The feasibility of operation of the reservoir on the basis of short- to medium-range water quality forecasts is not yet clear.
Comments
Session S5-01, Special Session: Computational Intelligence in Data Driven and Hybrid Models and Data Analysis I