Climate change is the main battle of our century. It is placed at the top of global priorities, pushing all levels of government to set a plan towards carbon neutrality. However, local governments often face a significant gap between policy ambition and the technical capacity required to support informed, data-driven decision-making. In the context of buildings, which are responsible for 40% of greenhouse gas (GHG) emissions, Urban Scale Energy Models (USEMs) have emerged as a promising solution to simulate energy use and CO₂ emissions across large building stocks.
This study contributes to the growing field of USEMs by proposing a statistical methodology for predicting CO₂ emissions at the urban scale, using the parish of Vila Franca de Xira, Portugal as a case study. The method uses Geographical Information Systems (GIS) and commonly available building attributes to develop a regression-based model focused on the residential sector. To give a taste of what a USEM could bring to public authorities, the results are visualized through an interactive and user-friendly online map.
The methodology was developed by testing multiple predictors, such as construction period and building type, to build a simplified model. While the statistical approach proves to be fast, replicable, and aligned with the technical skills of urban planners, it was found to be less accurate than machine learning-based models. Nonetheless, the research confirms the value of USEMs as a powerful decision-support tool, enabling local authorities to set more effective targets in the context of climate incentives.