Comparison of sensitivity analysis methods for pollutant degradation modelling: A case study from drinking water treatment Original Research Article

Five sensitivity analysis methods based on derivatives, screening, regression, variance decomposition and entropy are introduced, applied and compared for a model predicting micropollutant degradation in drinking water treatment. The sensitivity analysis objectives considered are factors prioritisation (detecting important factors), factors fixing (detecting non-influential factors) and factors mapping (detecting which factors are responsible for causing pollutant limit exceedances). It is shown how the applicability of methods changes in view of increasing interactions between model factors and increasing non-linearity between the model output and the model factors. A high correlation is observed between the indices obtained for the objectives factors prioritisation and factors mapping due to the positive skewness of the probability distributions of the predicted residual pollutant concentrations. The entropy-based method which uses the Kullback–Leibler divergence is found to be particularly suited when assessing pollutant limit exceedances.