EFFECT OF INPUT UNCERTAINTIES UPON SCENARIO PREDICTIONS FOR THE RIVERVECHT

Within the restoration plan for the river Vecht, the linked flow-water quality model DUFLOW has been applied to describe the behaviour of he avy metals. The model has been used to predict the effects of a number of scenarios for improvement of water quality. In this paper an analy sis of the input uncertainty and the effect upon the resulting uncerta inty in the output is presented. Both the influence of parameter uncer tainty and the uncertainty in the other input variables as: boundary c onditions loading from point and non-point, sources, and initial condi tions, has been studied. From reported ranges in literature an estimat e of the parameter uncertainty has been made. The uncertainty within t he other input variables has been assessed from available data. Throug h Monte Carlo simulation, using Latin Hypercube sampling the resulting uncertainty in the model predictions has been determined. From the un certainty analysis it appeared that it was not possible to discriminat e between the predicted impacts of some of the scenarios. Both the unc ertainty in the parameters and in the loading of the system contribute d to the overall model uncertainty, although their relative contributi on differed going from south to north in the system. Copyright (C) 199 6 IAWQ. Published by Elsevier Science Ltd