Lunch talk by Wouter Knoben

Title: “Estimation of non-stationary hydrological model parameters.”

Abstract for the presentation:

Recent research has shown that conceptual hydrological models decline in performance, when applied to periods with climate characteristics that are different from the period on which the model was calibrated. It has also been shown that calibrated parameter values show correlations with certain climate characteristics. Declined model performance as a result from different climatic conditions is an issue during climate change impact assessment. This study attempts to estimate, rather than calibrate, model parameter values from climate characteristics, in order to increase model performance under changing climatic conditions. The HBV model is used with the Polish Welna catchment as test case.

Correlations between optimal parameter values and certain climate characteristics are used as a basis for regression analysis. The resulting regression equations that estimate parameter values from climatic conditions are implemented in the HBV model. The regression model is compared to a base model to assess the difference between both models. Both models are then used for a climate change impact assessment to determine how the choice for regression of base model affects climate change impact assessment outcomes.