Modeling the spatiotemporal dynamics of electric power consumption in China from 2000 to 2020 based on multisource remote sensing data and machine learning

China's rising electricity power consumption (EPC) is strongly correlated with carbon emissions. Timely and accurate analysis of the spatiotemporal dynamics of EPC is important to the realization of carbon peaking and carbon neutrality goals in China. However, due to data quality problems and limitations of modeling methods, the estimation accuracy warrants further improvement. In this study, the EPC in China from 2000 to 2020 was estimated based on multisource remote sensing data using the Random Forest (RF) model. Compared with previous studies, the accuracy of this study was improved by 39%–47%. The reasons were combining multisource remote sensing data can mitigate the quality issues of nighttime light (NTL) data, and the RF can capture the nonlinear relations between remote sensing data and EPC. In addition, the spatial pattern of the average EPC in China was dominated by the low-level EPC, as well as showing an obvious increasing trend. We also found that in the middle reaches of the Yellow River and northern coastal China, the low-speed increase in EPC led to high carbon emissions and emission intensity. We suggest optimizing the fuel fix of energy and adjusting the industrial structure, combining them with scientific and rational spatial planning.