Recieved:

21/07/2026

Accepted:

10/09/2026

Page: 

–

doi:

http://dx.doi.org/10.17515/resm2026-1963vk0721rs

Views:

25

Identification of carbon emission mechanism typologies: A Data-Driven approach integrating GWR and SOM-FCM clustering

Yang Sun 1

1Manchester School of Architecture, The University of Manchester, Manchester, M1 7ED, England

Abstract

Spatial optimization of the built environment has become an important strategy to achieve carbon neutrality. However, most existing models ignore the spatial non-stationarity of driving forces. Therefore, this study combines GWR with an integrated SOM-FCM method for unraveling the local variations of built environment drivers of carbon emissions in downtown Guangzhou. The findings show that: (1) The geographical patterns demonstrate an obvious “south-west high and north-east low” emission trend with a clear core-periphery pattern in which intense emissions tend to be significantly concentrated in the downtown historical area. (2) Influencing factors have strong spatial variability; especially, road density and functional intensity aggravate emissions in the mature core, whereas mass transit and green space act as effective mitigating factors in the high-density southern areas. (3) The SOM-FCM framework synthesized these complex patterns into four distinct “carbon emission mechanism zones”, namely Vitality-Traffic Driven, Morphology Dominant, Location-Sensitive, and Industrial-Logistics, each characterized by unique profiles of built environment determinants. Ultimately, this research contributes a data-driven methodology for transitioning from uniform density regulations to zoning-based precision spatial governance in high-density metropolises.

Keywords

CO2 emissions; Built environment; Urban morphology; Geographically Weighted Regression (GWR); SOM-FCM clustering; Machine learning; Spatial heterogeneity; Precision spatial governance; Guangzhou; China

Cite this article as: 

Share This Article
LinkedIn
X
Facebook
journal cover
News & Upcoming Events