Non-parametric methods for global sensitivity analysis of model output with dependent inputs. (October 2015)

• Record Type: Journal Article
• Authors: Mara, Thierry A.; Tarantola, Stefano; Annoni, Paola
• Is Part Of: Environmental modelling & software. Volume 72(2015:Oct.)
• Page Start: 173
• Page End: 183
• Publication Date: 2015-10
• Subjects: Dependent inputs -- Rosenblatt transformation -- Variance-based sensitivity indices -- Dependent contributions -- Independent contributions -- Iman & Conover sampling procedure -- Radionuclide migration; 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.2015.07.010
• Languages: English
• ISSNs: 1364-8152

Abstract

Abstract: This paper addresses the issue of performing global sensitivity analysis of model output with dependent inputs. First, we define variance-based sensitivity indices that allow for distinguishing the independent contributions of the inputs to the response variance from their mutual dependent contributions. Then, two sampling strategies are proposed for their non-parametric, numerical estimation. This approach allows us to estimate the sensitivity indices not only for individual inputs but also for groups of inputs. After testing the accuracy of the non-parametric method on some analytical test functions, the approach is employed to assess the importance of dependent inputs on a computer model for the migration of radioactive substances in the geosphere. Highlights: We define a set of variance-based sensitivity indices for models with dependent inputs. The new sensitivity indices are those of the Rosenblatt transforms of the original variables. Non-parametric sampling-based strategies are proposed to compute the sensitivity indices. When input dependency is simply correlation, the simpler sampling procedure by Iman and Conover can be used. The proposed indices are computed and discussed for a radionuclide transport model a benchmark in sensitivity analysis.