AN IMPROVED FRAMEWORK FOR THE PARAMETERS REGIONALISATION OF HYDROLOGICAL MODEL

Regionalisation of hydrological model parameters is a simple approach to model ungauged basins. However, uncertainties in model parameters and catchment attributes hinder the regionalisation. In this context, this study proposes a methodology for modeling ungauged basins by pairing the regional model with the posterior distribution of parameters. The performance of regional models are evaluated by comparing the loss in performance and quantifying the uncertainty induced on the result of regionalisation. The study revealed the reduction in inconsistency among various regional models and improvement in performance while the results of regionalisation are constrained by regionalised ranges of parameters. Furthermore, the non-parametetric bootstrap methodology used to quantify of the uncertainty in regional models reveals that the indirect calibration method induced more uncertainty on the result of regionalisation compared to the conventional regionalisation schemes investigated in this study.