A Stochastic Water Quality Forecasting System for the Yiluo River

Abstract

A challenging problem for water quality management in northern Chinese rivers is their high loadings of organic pollutants and suspended solids, leading to complexities in producing effective water quality models. Also, uncertainties exist in many system parameters and their interrelationships. This study aims at developing a stochastic water-quality forecasting system and applying it to the Yiluo River, a tributary of the Yellow River with extremely high sediment and suspended-solid loadings. Extensive investigations of water quality in the river and the related pollution sources and watershed conditions were conducted. A one-dimensional BOD-DO model was developed to simulate water quality in the river, with interrelationships among water quality and the related source and sink conditions being explicated. A stochastic water-quality forecasting system was then developed to reflect random characteristics of many parameters, based on Kalman-filtering and self-adaptive techniques. The developed system was used for predicting DO and BOD levels in the Yiluo River. The results indicated that randomness in many system parameters and their interactions had been effectively handled; the accuracy of state estimation was generally satisfactory.

Keywords: Filter, forecast, Kalman, model, pollution, random, river, stochastic, uncertainty, water quality