REGIONALISATION OF POSTERIOR PROBABILITY DISTRIBUTION OF MODEL PARAMETERS : PREDECTION ON UNGAGUED BASIN

Abstract

In the case study presented here, the calibrated parameters of a Conceptual Rainfall-Runoff (CRR) model could not be uniquely identified. Moreover, the Posterior Probability distribution (PPD) of Model Parameters varied among basin depending upon the basin attributes and data aspects. So the assessment of parameter uncertainty of CRR models should become the integral part of regionalisation. For that reason, we proposed a regionalisation scheme which instead of regionalizing the single best value of model parameters, regionalizes the PPD of model parameters. Subsequently the number of model parameters are sampled from the estimated distribution and fed into the model. Regionalisation of PPD addresses the effect of model parameter uncertainty in the result of regionalisation. The uncertainties in model prediction quantified from proposed methods closely followed the prediction uncertainties quantified from PPD of model parameters conditioned on observation in the presented case study.