Researchers at the National University of Singapore have created an AI-based model that maps building carbon emissions in multiple cities, providing a tool to help develop more equitable and targeted climate policies.
Researchers at the National University of Singapore (NUS) have developed an open-source artificial intelligence model that accurately maps the carbon emissions of buildings across several major cities. The innovation promises to be a game-changer for policymakers aiming to devise targeted and equitable decarbonization strategies.
The model, born out of the College of Design and Engineering (CDE) at NUS, provides city planners with a detailed visualization of carbon emissions distribution and the driving factors behind them. This unprecedented clarity aims to help authorities craft smarter, fairer emission reduction strategies.
“Our model estimates operational carbon emissions of individual buildings at the scale of entire cities,” lead author Winston Yap, a doctoral student in CDE’s Department of Architecture, said in a news release.
Published in the journal Nature Sustainability, the research is spearheaded by Filip Biljecki, an assistant professor in the Department of Architecture.
The team’s model utilizes only open data, making it highly adaptable for various cities regardless of their data availability. It covers a range of urban environments, including Singapore, Melbourne, New York City (Manhattan), Seattle and Washington, D.C., mapping over half a million buildings in the process.
Remarkably, the model accounted for up to 78% of the variation in emissions across these diverse locales.
“Unlike previous approaches that rely on proprietary data, our open approach is designed to be transferable across cities, including those with different data availability conditions,” Yap added, underscoring the model’s versatility.
The study revealed critical insights into the spatial differences in urban emissions and identified influential factors such as urban form, planning history and income levels.
For instance, while tall buildings are generally more energy-efficient per unit area, dense urban areas may face higher cooling demands due to the urban heat island effect.
Surprisingly, suburban regions, dominated by low-rise detached homes, often rival city centers in total emissions contribution.
“Building emissions are not just about size or density; they’re deeply shaped by the unique context of each city, from its planning legacy to climate and economic conditions,” added Biljecki. “By using only open data, we’ve built a flexible framework that cities around the world can use to better understand their carbon footprint and plan more effective responses.”
The research also highlighted significant socioeconomic disparities.
Wealthier neighborhoods were frequently found to have higher per capita emissions. Notably, in Manhattan, a few large buildings accounted for more than half of the total building emissions.
This finding suggests that uniform carbon pricing or blanket regulations could unfairly impact lower-income communities struggling with older, less efficient infrastructure.
“Uniform carbon pricing or blanket regulations risk placing an unfair burden on lower-income communities that may already be struggling with older, less efficient infrastructure,” Biljecki added. “Our findings highlight the need for place-based strategies that take both emissions intensity and socioeconomic vulnerability into account.”
The model leverages diverse data sources, including satellite imagery, street view photos, population maps, road networks and local climate data, and uses advanced graph neural networks to capture spatial relationships between urban elements.
By making their approach entirely open-source, the researchers aim to bolster global efforts to cut emissions from the built environment and assist cities in meeting their climate targets.
“This work demonstrates the potential of open science and AI to accelerate urban sustainability,” added Biljecki. “It’s not just about understanding where emissions come from, but also ensuring that climate action is both effective and fair.”
Source: National University of Singapore