Constructing Generalized Local Carbon Emission Zones (Lcezs): A 2d/3d Urban Morphology Clustering Framework for Low-Carbon Planning

Abstract

The transition to low-carbon urbanisation is crucial for mitigating CO2 emissions. Previous research has characterised the low-carbon urban morphologies at the urban scale. However, there have been challenges in applying urban morphology control threshold values to achieve the goal of low CO2 emissions. This study introduces a novel approach by constructing Local carbon emissions zones (LCEZs) based on the general impact rules of urban form on residential and transport CO2 emissions (RTCE). After the downscaling process, geographically weighted regression was used to analyse the impact of 2D and 3D urban morphology factors on RTCE at a 123-meter resolution. Jenks Natural Breaks, K-means, and DBSCAN were applied to construct LCEZs. The Geodetecter model assessed the optimal cluster method and the representational performance of LCEZs, and their difference and consistency were tested and validated. The results showed that the degree of spatial congestion was the most significant influence of the seven urban morphology factors on RTCE. The impact of road length was the least. The Natural break method emerged as the most effective clustering technique for LCEZs construction. Building coverge ratio, Mean building height, and Road length were the most representative factors for RTCE in urban centres. Additionally, we analysed the variations and consistencies among LCEZs across 24 urban centres to validate the construct of LCEZs. Finally, suggestions for low-carbon urban construction were proposed based on the proportion of RTCE cold spots in different functional areas across various LCEZs.

Keywords: Cities, Urban morphology, Climate change mitigation, machine learning, GWR, GeoDetector