Critical processes-based parameter uncertainty analysis in whole building life cycle assessments

Abstract

Uncertainties are inherently present in any life cycle assessment (LCA) but seldom explicitly declared, so deterministic result communication prevails, despite their potential to mislead interpretation. Monte Carlo simulation is the probabilistic method most frequently used for uncertainty analysis. As it involves extensive data amounts, prioritizing critical processes suits particularly the case of evaluations composed of numerous processes, such as those of whole buildings (wbLCA). As a proof of concept, we developed a protocol and illustrated its application to a wbLCA study developed in a - foreground and background - data scarcity context The four-step protocol comprises contribution analysis to select the higher impact (critical) processes; semi-quantitative analysis through the Pedigree matrix; quantitative analysis with the Monte Carlo simulation; and results visualization using a parameter categorization system. The database processes had high parameter uncertainty, strongly influenced by the extent that such uncertainty is informed within the databases used. The protocol allowed identifying the most critical (high contribution and high uncertainty, simultaneously) processes in the case studied and respective parameter uncertainty estimates. The established factors in the Pedigree matrix are little sensitive to contextual contrasts. Hence, using the matrix for characterizing data quality demands unreasonable intensive work for subtle changes in the uncertainty profile. Thus, we suggest a revision of current coefficients to ensure appropriate data quality loss representation and to automate contextualization routines. The parameter categorization system promotes a dialogue between analysts and designers, fundamental for environmentally sound decision making.