Key points are not available for this paper at this time.
ABSTRACT There is an increasing demand for carbon emission data accuracy, but existing studies rely mainly on single or limited economic and social indicators such as nighttime lighting and population density, while largely neglecting the influence of environmental factors, resulting in insufficient carbon emission data accuracy. To address this limitation, and considering the direct and indirect impacts of environmental variables on carbon emissions, we introduced land cover data as an environmental factor and developed a Socio‐Economic‐Environmental (SEE) downscaling framework. The framework integrates population, gross domestic product (GDP), land use data, and NDVI to capture the interactions between natural and artificial environmental factors. Furthermore, by systematically comparing different combinations of variables as well as alternative modeling approaches, we identified the optimal model configuration for high‐resolution carbon emission estimation. After applying the model to the Beijing–Tianjin–Hebei (BTH) and Yangtze River Delta (YRD) urban agglomerations, we successfully downscaled carbon emission grids from a resolution of 0.1° to 1 km. Through the comparative analysis of the 1‐km resolution carbon emission grid data in 2019 and 2020, we reveal the differentiated impacts of the natural and artificial environments on carbon emissions, as well as the refined distribution characteristics of carbon emissions within different urban agglomerations. By integrating ecological and anthropogenic factors, the SEE model covers a wider range of carbon emission determinants and improves the precision and accuracy of carbon emission data.
Shen et al. (Fri,) studied this question.