Dissecting uncertainty in life cycle assessment studies for sustainable pavement management

The integration of environmental performance into pavement management (PM) is complicated by uncertainties that often go unaddressed in traditional life cycle assessment (LCA) studies. These uncertainties are particularly acute in the early stages of PM, where data is limited. To address this gap, our study introduces a
refined uncertainty analysis methodology focusing on improved sensitivity analysis within LCA, targeting the evaluation of environmental impacts of maintenance and rehabilitation (M&R) measures. Our framework accounts for various types and sources of uncertainty, including the variability of input parameters and the choices made in methodology. We apply three global sensitivity analysis (GSA) methods—Extra Trees, Sobol, and PAWN—to identify key variables and evaluate their influence on uncertainty. Employing a case study of the Dutch main road network, we assess the environmental impacts related to global warming and ozone layer depletion across multiple reclaimed asphalt pavement (RAP) scenarios. Our results underscore the significant role of pavement-vehicle interaction (PVI) in driving environmental impacts and uncertainty while highlighting transportation as a consistent source of uncertainty. Notably, Extra Trees and PAWN emerge as strong, computationally efficient alternatives to the more commonly used Sobol method. Overall, this study offers actionable insights for PM, enabling targeted improvements in environmental performance and providing a nuanced understanding of how different GSA methods can effectively manage uncertainties.