Influence of rating curve uncertainty on daily rainfall–runoff model predictions.

River discharge observations are usually affected by uncertainty, which is due to many concurrent causes and strongly affects the response of rainfall-runoff models. The present paper is aimed at studying the influence of imperfect rating curve knowledge on the uncertainty of the response of a daily conceptual linear-nonlinear rainfall-runoff model. To describe the impact of imperfect rating curve knowledge, simulations have been conducted using a conceptual rainfall-runoff model and continuous daily series of rainfall, air temperature and discharges recorded in a Sicilian catchment. The GLUE procedure was used to introduce the uncertainty of the rating curve in a classical rainfall-runoff model uncertainty analysis by randomly modifying the original rating curve. The final results show an increase of uncertainty with respect to the original rating curves for the majority of model predictions, but more significantly for overestimated rating curves than underestimated ones