Uncertainty Analysis in Water Quality Modelling: Case Study: Hjarbæk Fjord

An important part of regional planning of water resources and quality is efficient design of monitoring systems and proper use of hydrologic models (Beven 1993). In the design of monitoring systems as well as validation of numerical models, based on, for example, the equation of continuity such as hydrologic routing models and mass balance nutrient models, it is essential to estimate the uncertainties of the model-predictions. This paper presents an implementation of a first-order analysis for estimating the error-propagation when introducing mass balance models as to predict nutrient-concentrations. The uncertainty assessment, developed from a first order theory, is implemented in the analysis and modelling of Hjarbaek fjord in Denmark. The project includes hydrological modelling of input of water and nutrients to the fjord from tributaries, and a hydrodynamic estimation of water levels and velocities in the fjord. A two-system water quality box-model is used for estimation of concentrations in water and sediment phases. The system uncertainties are analysed, starting with input data uncertainties and the error propagation to the final concentration estimates, in order to optimise the future monitoring programme, and to control the model results.