Effects of rating-curve uncertainty on the calibration of numerical hydraulic models

This paper analyses the uncertainty of rating curves and how this uncertainty affects the calibration of numerical hydraulic models. Rating curves, relating stage and flow discharge, are traditionally used for describing boundary conditions. The uncertainty associated with rating curves is often neglected, and generally considered to be less important than other factors. We performed a series of simulation experiments aimed at: (1) quantitatively assessing the uncertainty of the curves; (2) investigating the effects of uncertainty on the calibration of Manning’s roughness coefficient. We used a quasi-twodimensional (quasi-2D) model of the middle-lower reach of the River Po (Northern Italy) to simulate 10 different historical flood events for the hydrometric river cross-section located in Cremona. Using the simulated data, we mimicked 5 measurement campaigns for each flood event and we corrupted the synthetic discharge data values of each campaign according to the indications on measurement errors reported in the literature. For each synthetic campaign, we applied different procedures for rating curve estimation, and we quantified the 90% confidence interval of the estimated curves. Then, we performed an additional set of model runs downstream of the Cremona’s cross-section to assess how the uncertainty of rating curves affects the estimated Manning’s coefficients during the calibration phase. The results of the study are analysed and discussed relative to the variability of Manning’s coefficient reported in the literature for lower reaches of large natural streams.