Life cycle assessment of urban road networks: Quantifying carbon footprints and forecasting future material stocks

This study introduces a comprehensive framework designed to assess the carbon footprint of urban road networks, covering all stages from production to maintenance. Complementing this framework, a dedicated prediction model is also developed. Utilizing a machine learning methodology, specifically the Gradient Boosting Decision Tree (GBDT) algorithm, this model aims to accurately predict the material stock of urban road networks, thereby enhancing the overall understanding of their environmental impact. A case study focusing on Nanjing's infrastructure revealed that material production constitutes 78 % of the total emissions. Additionally, it was observed that the growth in the carbon footprint of Nanjing's urban road network has slowed down in recent years. The machine learning-based model, analyzing data from 2014 to 2020, accurately predicted the total material stock for 2021 and the stocks of different road construction materials, with a relative error under 2 %. This approach aids in the design of carbon-neutral urban transport systems and represents a stride in combating climate change. [Display omitted] •Developing a framework to assess the material flow of urban road networks.•Assessing urban road networks' environmental impact via carbon footprints.•Introduced machine learning method for urban road material stock prediction.•Achieved a predictive model with less than 2 % deviation in material stock.