Sensitivity analysis: a statistical tool for a green roof model analysis
Résumé
Green roofs are sustainable solutions to manage water runoff from rain events in urban areas. They allow to decrease the peak flow from 22% to 93% in water system and delay it by 0 to 30 min [Li and Badcock Jr (2014)]. The overall performances are directly associated to the green roof characteristics (substrate retention, dimension, type of vegetation, etc.). In order to better understand the hydrological dynamics, the water retention capacity of the substrate is often modelled by the Van Genuchten – Mualem equations coded in the software Hydrus-1D© [Simunek et al. (2008)]. However, the model parameters, as soil parameters or vegetation variables, are challenging to determine as they are difficult to measure accurately through experiments.
Thus, the uncertainties of these parameters are propagated to the water retention capacity simulated and need to be analysed to calibrate the model. Global Sensitivity Analysis (GSA) is a statistical tool which allows to analyse these uncertainties. Global sensitivity analysis can be used to verify and better understand the model behavior, help for calibration and target parameters to focus on. Based on variance decomposition of model output, GSA quantifies effects of parameter uncertainties on the output [Saltelli et al. (2008)]. It explains which parameter variations are responsible for the output variations by highlighting both the influential and the not-influential parameters. In this study, global sensitivity analysis is used to calibrate the model and quantified the influence of the soil and vegetation parameters on the water retention capacity over time in-situ data, boundary conditions, meteorological data come from an in-situ experimental green roof platform located in Tomblaine (France).
This work is part of a collaboration between Cerema and the Research Center for Automatic Control
Origine | Fichiers produits par l'(les) auteur(s) |
---|