Abstract: Objective　With the advancement of global climate governance and China's "dual carbon" goals, urban agglomerations have become key spatial units for carbon emission reduction. The Hohhot-Baotou-Ordos-Yulin urban agglomeration, as an important energy base and ecological barrier in northern China, is dominated by resource-based industries, featuring large total carbon emissions, significant regional disparities, as well as the increasing pressures of ecological constraints and transformation. Existing studies on carbon emissions mostly focus on a single scale, lacking cross-scale comprehensive analysis from urban agglomeration to city and county levels, making it difficult to accurately reveal spatiotemporal 453 differentiation patterns and transmission mechanisms. To support regional coordinated emission reduction and green transformation, there is an urgent need to conduct research on the spatiotemporal characteristics of carbon emissions from a cross-scale perspective, so as to provide a scientific basis for formulating differentiated low-carbon policies. Methods　This study employs the standard deviation ellipse method to visually reveal the spatial distribution patterns and evolution trends of energy consumption-related carbon emissions by comparing the coverage, spatial orientation, and dispersion levels of ellipses across different years. Meanwhile, exploratory spatial data analysis (ESDA) conducted on a GIS platform is utilized to decipher the spatial correlation characteristics and distribution patterns of carbon emissions within the study area from 2012 to 2022. Furthermore, a Markov chain model is employed to investigate the dynamic shifts and evolutionary patterns of carbon emission levels within the Hohhot-Baotou-Ordos-Yulin urban agglomeration. Results and Discussion　From 2012 to 2022, the overall carbon emissions from energy consumption in the Hohhot-Baotou- Ordos-Yulin urban agglomeration exhibited a phased pattern of initial increase followed by decline. Spatially, high-emission core cities were concentrated in Ordos and Yulin. The major axis of the standard deviation ellipse consistently oriented southwest to northeast, with the coverage area at the municipal level tending to shrink while the coverage area at the county level expanding accordingly. Regarding spatial clustering, the agglomeration was dominated by two types, namely high-high (HH) and low-low (LL). This pattern was primarily shaped by the high-carbon cluster formed by Fugu County and Shenmu City, and the low-carbon cluster formed by Jia County, Mizhi County, and Suide County. From a dynamic transfer perspective, the influence of spatial neighborhood effects exhibits heterogeneity, that is higher-level neighborhood environments significantly promote growth in regions with moderate and equivalent carbon emission levels, but have a negligible impact on low-level regions. Conversely, the radiating and driving effects of high carbon emission levels themselves are more pronounced in low-level regions. Conclusion　This paper systematically analyzes the temporal changes, spatial patterns, and dynamic transfer mechanisms of carbon emissions within the Hohhot-Baotou-Ordos-Yulin urban agglomeration at the municipal and county/district levels. It holds significant theoretical and practical implications for optimizing urban spatial structures, deciphering the spatiotemporal distribution of economic-geographical factors, and advancing sustainable urban development for this urban agglomeration.

Key words: Hohhot-Baotou-Ordos-Yulin urban agglomeration, night light, carbon emissions, space-time evolution