(Universidad de Cuenca, 2011-07) Vázquez Zambrano, Raúl Fernando; DIUC; Universidad de Cuenca; Dirección de Investigación de la Universidad de Cuenca
The analysis of the most commonly used measures of hydrological/hydraulic model performance was herein carried out by means of their statistical examination and illustrative modelling applications. In doing so, the model performance indexes were classified in two groups, according to the type of error (or residual) that those indexes are measuring: (i) statistics measuring the average systematic error (model bias); and (ii) statistics measuring the average combined systematic and random discrepancies among simulations and observations. The reader can in this way easily select a set of unrelated statistics to report on model performance. The manuscript addresses as well the main pitfalls of some of the most popular statistics used in scientific literature and suggests some approaches to overcome such potential pitfalls when addressing model performance.
