Estimating hyperlocal traffic CO₂ by customizing spatial relationship: An analysis from Digital Footprint data

Abstract. Globally, transport accounts for around one- fifth of CO2 emissions. However, , leveraging the DF data in modeling hyperlocal traffic CO2 and exploring the potential environmental justice is still underexplored. Here, we first extract traffic flows from the DF data, including individual GPS tracks, traffic counts, and car ownership rates in Glasgow, UK, then redefine the spatial relationship by incorporating traffic flows into the Spatial Weight Matrix (SWM), and finally predict the hyperlocal traffic CO2 based on customized SWM. We find that, compared to traditional distance-based SWM, incorporating the real traffic flows into the SWM could better predict hyperlocal traffic emissions, with the Manski model performing best (R2 = 0.62). Besides, the Manski model shows that income and car ownership rates are dominant factors related to traffic CO2. Based on this, we reveal two aspects of environmental justice: 1). Distribution inequality - the high-income areas also have higher levels of car ownership rates, indicating higher barriers and challenges for low-income communities; 2). Contributor inequality - most high traffic CO2 emissions are produced by nearby affluent areas with high car ownership rates, whereas the low-income areas suffer more traffic emissions produced by them, which indicates that disadvantaged groups bear the costs of emissions disproportionately generated by the advantaged. This pilot study explores the application of DF data in environmental monitoring, carbon justice, and climate mitigation to create an equitable and sustainable living environment.