Sensitivity and uncertainty analyses of retention performance of green roofs considering aleatory and epistemic uncertainties

Abstract

A methodological framework is presented to integrate epistemic uncertainties into an aleatory-based probabilistic green roof model. In doing so, the sensitivity and uncertainty of the mean and standard deviation of retention ratio are assessed. Regression-based global sensitivity analysis showed that the first-order effect of model parameters plays the dominant role over the higher-order effect in contributing the overall uncertainty of statistical moments of retention ratio from the probabilistic green roof model used. Inclusion of epistemic uncertainties allows comprehensive evaluation of uncertainty associated with the performance of a green roof system, which further facilitates risk-based design of green roof systems.