Combined analysis of time-varying sensitivity and identifiability indices to diagnose the response of a complex environmental model. (February 2017)

• Record Type: Journal Article
• Authors: Ghasemizade, Mehdi; Baroni, Gabriele; Abbaspour, Karim; Schirmer, Mario
• Is Part Of: Environmental modelling & software. Volume 88(2017)
• Page Start: 22
• Page End: 34
• Publication Date: 2017-02
• Subjects: Temporal sensitivity -- Identifiability -- Preferential flow -- HydroGeoSphere -- Output uncertainty; Environmental monitoring -- Computer programs -- Periodicals; Ecology -- Computer simulation -- Periodicals; Digital computer simulation -- Periodicals; Computer software -- Periodicals; Environmental Monitoring -- Periodicals; Computer Simulation -- Periodicals; Environnement -- Surveillance -- Logiciels -- Périodiques; Écologie -- Simulation, Méthodes de -- Périodiques; Simulation par ordinateur -- Périodiques; Logiciels -- Périodiques; Computer software; Digital computer simulation; Ecology -- Computer simulation; Environmental monitoring -- Computer programs; Periodicals; Electronic journals; 363.70015118
• DOI: 10.1016/j.envsoft.2016.10.011
• Languages: English
• ISSNs: 1364-8152

Abstract

Abstract: Sensitivity and identifiability analyses are common diagnostic tools to address over-parametrization in complex environmental models, but a combined application of the two analyses is rarely conducted. In this study, we performed a temporal global sensitivity analysis using the variance-based method of Sobol' and a temporal identifiability analysis of model parameters using the dynamic identifiability method (DYNIA). We discuss the relationship between the two analyses with a focus on parameter identification and output uncertainty reduction. The hydrological model HydroGeoSphere was used to simulate daily evapotranspiration, water content, and seepage at the lysimeter scale. We found that identifiability of a parameter does not necessarily reduce output uncertainty. It was also found that the information from the main and total effects (main Sobol' sensitivity indices) is required to allow uncertainty reduction in the model output. Overall, the study highlights the role of combined temporal diagnostic tools for improving our understanding of model behavior. Highlights: Main and total effects are both required to identify the important parameters. Identifiability is necessary but not sufficient for uncertainty reduction. Soil moisture affects temporal variability of sensitivity and identifiability. Combined diagnostic tools improve our understanding of complex model behavior.