Análisis de la distribución estadística del coeficiente de rugosidad y su efecto usando la metodología GLUE en tres morfologías de ríos de montaña

Abstract

The estimation of the roughness coefficient is of vital importance in one-dimensional hydrodynamic models. There are methods to estimate the roughness parameter in a hydrodynamic model based on sampling the possible values and comparing them with field measurements. The statistical distribution assumed in these samples is a uniform one where all data in the range are equally probable. The GLUE method is one such method that uses a uniform distribution. In this methodology, the observed conditions are reproduced by combinations of models, parameters and variables according to the concepts of equifinality and likelihood. It is known that: the common morphologies of natural channels in mountain rivers are: flat bed, cascade and gradient. The main aspect of this qualification work is based on determining the influence of changing from assuming the uniform distribution as the default for roughness parameters to a statistical distribution according to roughness measurements made in the different case studies. The distribution determined by a previous statistical analysis was used in a GLUE test, and its efficiency was compared with tests based on a uniform distribution. Finally, the convergence was analyzed according to the number of iterations required and the significance level of the results, determining the false claim of obtaining better accuracy and speed of convergence thanks to the change of statistical distribution used.