Automating carbon risk assessment of residential buildings using scalable energy and emissions data in city information models

The residential building sector accounts for approximately 17-20% of global CO₂ emissions, making it a critical target for decarbonization efforts aligned with the Paris Agreement's 1.5°C warming limit. This study presents an automated framework for carbon risk assessment that integrates scalable energy and emissions data into City Information Models (CIMs), addressing the limitations of manual assessment tools that require extensive data input and lack scalability. The framework leverages TABULA building archetypes and power law scaling (optimal exponent b=0.85) to extrapolate energy demand and CO₂ emissions from minimal input data such as building geometry, address, heating system, and renovation status, thereby enriching CIMs with dynamic sustainability metrics. Validation using 10 Bavarian residential buildings demonstrated robust predictive accuracy, with energy demand predictions achieving a Normalized Mean Bias Error (NMBE) of 4.77% and CO₂ emissions an NMBE of 4.17%, both within ASHRAE Guideline 14 thresholds (±5%). Strong correlations were observed (Pearson's r=0.825 for energy, 0.774 for CO₂), though stranding year predictions under the CRREM 1.5°C energy reduction pathway showed deviations of up to four years, reflecting limitations from static decarbonization assumptions. This also highlights the framework’s intended use as a portfolio-scale screening tool for relative transition risk, rather than as a deterministic predictor of asset-level compliance dates. The framework quantifies excess carbon costs using EU ETS price trajectories, offering policymakers a scalable tool for prioritizing retrofits and mitigating stranded asset risks. Future enhancements should integrate dynamic grid decarbonization, machine learning for occupancy modeling, and expanded validation across diverse climatic contexts to strengthen applicability for global sustainability initiatives.