As the largest carbon emitter in the world, China is facing increasing challenge to reduce CO₂ emissions. Given this issue, exploring the influencing factors is of great significance for scientific low-carbon emission policymaking. Although previous literature has explored the effects of urbanization on CO₂ emissions, the impact of the space of flow on urban carbon emissions have been less explored. Due to the increasing connection between cities, its impact on urban carbon emissions cannot be ignored. Thus, this paper takes the space of flows into account as an aspect of urbanization to supplement the existing literature and empirically examines the multiple effects of urbanization on CO₂ emissions in the Pearl River Delta (PRD) urban agglomeration. By using a STIRPAT model, statistical data, and web crawler data, we examined impacts of different types of urbanization on CO₂ emissions. Our empirical results show that: (1) Within the PRD urban agglomeration, urban linkage intensity is strongly connected to urban socioeconomic growth, establishing a geographical structure with Guangzhou and Shenzhen as the double core. (2) Our results show that urbanization exerts two opposite effects on CO₂ emissions: positively connects carbon emissions with population urbanization, integrated urban linkage flow, and energy intensity, whereas economic urbanization and social urbanization are shown to be negatively correlated. However, spatial urbanization has no significant positive effect on urban CO₂ emissions. (3) It is worth noting that urban linkage flows are the second most important factor affecting urban carbon emissions after economic urbanization. Our study could formulate effective planning suggestions for future CO₂ emission reduction paths and development modes in the PRD.

Building carbon emissions significantly drive climate change, yet obtaining high-resolution macro-scale data reflecting spatial variations remains a challenge. This paper proposes a novel method for calculating urban-scale carbon emissions based on building stock and big data mining. By inferring building ages using Points-of-Interest (POI) and the K-Nearest Neighbours (KNN) algorithm, and determining building types through integration with building vector and land use, the study constructs a high-resolution building carbon emissions map. The findings as follows: (1) Guangzhou’s total building carbon emissions reached 1.103 billion tons. Carbon emissions were predominantly concentrated in the operational phase (56.2%). Residential, industrial, and commercial structures contributed significantly. (2) The carbon emissions exhibited a spatial pattern characterized by high-value concentration in the central urban area and low-value dispersion in peripheral zones. (3) Buildings constructed between 2000 and 2004 exhibit the highest carbon emissions, with per-unit emissions gradually increasing over time. Carbon emission hotspots demonstrate a dynamic temporal migration pattern of “expanding outward from the centre”. (4) The potential for building demolition and material recycling is projected to peak in the 2030s, occurring earlier in core urban areas than in peripheral new districts. (5) Spatial coupling between building carbon emissions and population distribution is stronger than with GDP. Central urban areas exhibit efficient aggregation, while peripheral new districts show inefficient decoupling. This study provides quantitative evidence and methodological references for predicting and regulating building carbon emissions at the urban scale.