Research On Parameters’ Uncertainty In QUAL2K Water Quality Model

Faced with the grim situation of water resources, to take appropriate measures to protect, manage and improve water resources is very necessary. Management decisions are usually enacted according to water quality evaluation and prediction results. Therefore, the water quality evaluation and prediction results are accurate or not has an absolute impact on the effectiveness of management decisions.Water quality model is an effective means for water quality modeling. Uncertain factors inevitably exist in all stages of constructing a water quality model. The simulation results are affected by uncertain factors. Analysis of uncertain factors will help improving the reliability of the model results, so that the simulation results can better reflect the real situation of water quality. The uncertainty of water quality model can be divided into the uncertainty caused by model structure, the uncertainty of input data and the uncertainty of parameters. The uncertainty caused by model structure is the fundamental source of the uncertainty of water quality model. Due to the lack of in-depth study on theoretical basis and effective techniques of the uncertainty caused by model structure, it is often analyzed by identifying the statistical law of parameters. Thus, to analyze the uncertainty of parameters is particularly important.QUAL2K is a suitable model to simulate rivers with tributaries. CBODf and CBODS are used to describe organic matter in this model. This paper studies the relationship between BOD5, CBODf and CBODU, the results shows that without considering the impact of external factors on dissolved organic matter, BOD5 can be obtained through the calculation of the simulation result of CBODf by QUAL2K.In this paper, the Bayesian theory and MCMC method are combined to propose an effective method to study the uncertainty of parameters in QUAL2K. Taking a section of Boulder Creek River as the research object, the value of parameters affecting CBODf are determined by using Bayes-MCMC. And the feasibility and validity of this method is verified by calculating. The sensitivity of parameters and its influence on model’s output can be obtained by study the uncertainty of model parameters. To study the uncertainty of model parameters can help to improve the reliability of QUAL2K.