Integrating planetary boundaries into energy system optimisation models for absolute environmental sustainability  assessment:  a methodological framework

Transforming energy systems to address climate change must avoid shifting burdens to other planetary boundaries. Yet most energy models still focus primarily on greenhouse gas reduction, often neglecting broader life-cycle impacts on critical Earth-system processes. This narrow scope can unintentionally shift environmental pressures from climate to land, materials, or ecosystems, even in scenarios deemed sustainable. To address this gap, we present a modeling framework that integrates prospective life cycle assessment and planetary boundary metrics into a national energy optimization model. The framework first integrates life-cycle and planetary-boundary metrics into a multi-objective optimization model, then uses goal programming to characterize the different trade-offs in environmental impacts, and finally applies interpretable machine learning to streamline the overwhelming set of trade-offs into system archetypes. Applied to a 2045 low-carbon scenario for Belgium as a case study, the approach reveals that all near-optimal pathways meet the climate target but exceed boundaries for ecotoxicity and particulate matter under the most permissive downscaling principle, indicating that decarbonization alone is insufficient for absolute sustainability. From the near-optimum space exploration and interpretable machine learning, four distinct system configurations emerge-dominated respectively by electrification, hydrogen, renewable power capacity, or mixed strategies-each achieving similar costs yet distinct environmental trade-offs. Future extensions of this work should incorporate dynamic demand changes, and behavioral adaptations to further align energy systems with absolute sustainability targets.