Â鶹ÒùÔº

Carbon emissions continue to increase at record levels, fueling climate instability and worsening air quality conditions for billions in cities worldwide. Yet despite global commitments to carbon neutrality, urban policymakers still struggle to implement effective mitigation strategies at the city scale.

Now, researchers at Notre Dame's School of Architecture, the College of Engineering and the Lucy Family Institute for Data & Society are working to reduce carbon emissions through advanced simulations and a novel artificial intelligence-driven tool, EcoSphere.

"Our goal is to develop tools that assess the carbon emission reduction and mitigation potentials of the built environment infrastructure—both through renovation and improved new construction. At the city scale, such a tool can offer data on building components and lifespans to support decision-making by policymakers and city planners," said Ming Hu, the associate dean for research, scholarship and creative work in Notre Dame's School of Architecture.

Hu studies how embodied carbon can be analyzed to develop greener cities. Often considered a blind spot in urban sustainability due to limited standardized data, embodied carbon includes the carbon dioxide emitted during the construction of buildings. It accounts for almost 40% of energy-related COâ‚‚ emissions.

Drawing from the life cycle assessments and renovation rates of more than 1 million buildings in Chicago, Hu worked with Siavash Ghorbany, a Notre Dame doctoral student in civil and environmental engineering, to develop a simulation of real-world urban dynamics. The model and the tool both help identify future mitigation strategies for reducing carbon emissions.

The findings are in the journal npj Urban Sustainability.

Hu and Ghorbany employed a "bottom-up" approach to tackling the project, calculating urban emissions by aggregating detailed data from individual buildings, including materials, age and structural characteristics.

The research generated over 350,000 simulated scenarios and revealed that strategies focused on renovation and extending building life can significantly reduce embodied carbon emissions. The results showed that new construction produces up to 7,500 times more COâ‚‚.

Renovation, Hu suggested, is often a more sustainable option.

"While new construction introduces greater uncertainty in emission outcomes, our findings show that strategic building updates—whether through renovation or carefully planned new development—can significantly mitigate these risks. However, an increase in building size can offset potential carbon savings, underscoring the importance of urban planning approaches that prioritize renovation, preservation and efficiency where possible."

For city planners and local policymakers, having access to this data in a user-friendly way may generate more effective advocacy for policies that can work to slow the rate of carbon emissions in cities.

Notre Dame researchers Matthew Sisk, codirector of the Civic-Geospatial Analysis and Learning Lab and associate professor of the practice in the Lucy Family Institute for Data & Society, Chaoli Wang, professor of computer science and engineering and Siyuan Yao, postdoctoral candidate in computer science and engineering, teamed up with Hu and Ghorbany to transform the scenario-based simulations into an AI-powered platform called EcoSphere, with results in Automation in Construction.

EcoSphere integrates national building datasets with embodied carbon data, Google Street View, satellite imagery and advanced machine learning techniques to generate graphics that can help city planners and non-experts visualize emissions data.

EcoSphere integrates national building datasets with embodied carbon data, Google Street View, satellite imagery and advanced machine learning techniques to generate readily available graphics that can help city planners and non-experts visualize emissions data. The EcoSphere interface provides information in a visualization dashboard, with the ability to dive deeper into simulation outcomes to understand potential cost implications, variables that drive simulation scenarios, the impact of mitigation strategies and cost and emission comparisons.

To evaluate the tool's effectiveness, case studies were carried out in Chicago and Indianapolis. In both cities, EcoSphere demonstrated how varying construction methods and policy decisions can significantly impact a city's carbon footprint and economic costs.

"EcoSphere uses machine learning not just to process these large datasets and imagery—but to understand it," Sisk said. "By combining computer vision, geospatial analysis and large language models, we can generate detailed carbon profiles in real-time for entire cities, making sustainable urban planning faster, smarter and more accessible."

Beyond city planning, EcoSphere has wider applications for use in school systems as a teaching tool for students to explore how carbon impacts the environment. Additionally, it can be integrated with smart city and digital twin platforms for real-time decision-making and monitoring. Governments can use it to forecast the long-term effects of policy choices and to craft more effective carbon reduction regulations.

Hu is hopeful that the tools will provide a positive impact on cities in the US, where robust, data-driven strategies can work toward neutralizing carbon emissions.

"Together, these studies show how detailed data and smart software can help empower city planners to make informed decisions for a greener future," Hu said.