Robust policy evaluation in material flow analysis under deep uncertainty

Abstract

This study proposes a novel framework for robust policy evaluation in material flow analysis (MFA) under deep uncertainty. It is motivated by our observation that, under data scarcity, conventional uncertainty analysis methods may overstate the available knowledge by attempting to characterize uncertainty beyond what limited empirical evidence can support. We propose a shift away from identifying the most probable policy outcomes and their associated uncertainty, toward identifying policies that remain reliable despite uncertainty. This is achieved by integrating MFA with exploratory modeling and analysis, which (i) characterizes broad, plausible uncertainty ranges to incorporate extreme scenarios and (ii) assesses policy performance across multiple robustness metrics to reflect diverse perspectives. In addition, the study presents methodological suggestions for handling the sum-to-one constraint of transfer coefficients, including a constrained elicitation procedure for defining uncertainty bounds, a sampling approach that preserves the sum-to-one property without requiring normalization, and a vulnerability analysis method that remains applicable despite this constraint. An illustrative case study demonstrates the operationalization of the framework and highlights its strengths and limitations under simplified conditions. Further research is needed to refine the methodology and enable a systematic application in practice.