ABSTRACT
Life cycle assessments (LCAs) typically use deterministic methods, yielding single-value results without accounting for uncertainty in life cycle inventory data. Existing approaches to uncertainty modeling in LCA often lack transparency, accuracy, and practical utility for practitioners. This study refines a novel uncertainty modeling method by employing kernel density estimation (KDE) with variable weighting to reflect production quantities by manufacturing country and variable bandwidths to capture product-level uncertainties in environmental product declarations (EPDs). Using 140 structural steel EPDs and global production data from World Steel, this study demonstrates how incorporating product-specific uncertainty and production volume can influence embodied carbon coefficient (ECC) uncertainty models. Weighting each EPD relative to its country’s share of global steel production ensures a balanced representation. This improved uncertainty modeling approach enables designers to demand greater data transparency from material suppliers. Although the availability of digitized EPDs is limited, legislative trends indicate a likely increase in production and access for EPDs of standard building materials.
