<p>Green roofs have become a key component of sustainable urban development. The hydrological performance of a green roof is characterized by the water content within the different layers of the roof. The models describing the water content involve soil, vegetation, and weather parameters, which may be difficult to determine and are therefore considered uncertain. It is desirable to quantify the uncertainty associated with green roof performance in order to enable a more comprehensive assessment or design. This paper provides a methodology for conducting sensitivity analysis of a complex model involving time-varying uncertain parameters and a time-varying output. It focuses on an original approach based on statistical analysis of the data and the Iman and Conover procedure to generate consistent random samples that are independent but follow a prescribed time correlation matrix. Furthermore, sensitivity is computed at each time step to better understand the behavior of the green roof according to seasonal changes and various rainfall events. This study highlights the parameters that influence water content during dry and wet periods over the course of one year.</p>