Mapping agency in urban metabolism research. A systematic characterisation of urban-level resource-use assessment models based on levels of engagement with agents

• A review of 1858 urban metabolism studies were retrieved concerning the modelling of human agency. • Three disciplines, social ecology, urban ecology , and industrial ecology are involved. • A model characterization is provided to investigate the models’ engagement with agency. • Variables are agency-defining terms and level of engagement with agents modelling. • Material Flow Analysis is the most comprehensive model with respect to the variables. For several years, industrial, social, and urban ecology have contributed to developing, testing, and validating city-scale modelling tools that help understand society-nature relationships through energy and material exchanges. These tools can track and represent metabolic processes from different perspectives while handling increasingly complex datasets. However, few studies have examined how urban metabolism models incorporate different forms of agency deployed by actors in resource management. Additionally, reviews comparing modelling approaches across disciplines in urban metabolism research remain limited. To address these gaps, this study explores whether and to what extent urban-level modelling traditions in industrial, social, and urban ecology can systematically represent how agents influence metabolic stocks and flows. First, a systematic literature review of urban metabolism studies focusing on computational and mathematical models is conducted across these three disciplines. Second, models are compared based on how they engage with the concept of “agency” and represent metabolic agents in the modelling process. Finally, a systematic characterisation of these models is provided, highlighting their differences and complementarities in addressing agency. The results show that MFA is the most comprehensive model, applied across all disciplines and engagement levels with nine different terms for “agents,” whereas SD, WB, CA, and EFA are the least comprehensive, limited to a single discipline and one agent term at one level. This characterisation provides a foundation for developing integrated decision-making frameworks for selecting and coupling models in future urban metabolism research.