Forecasting China’s CO2 emissions and identifying key drivers: an application of the improved RFAGM model and LMDI decomposition methods

Abstract

ABSTRACT China, the world’s largest CO2 emitter, has pledged to reduce its carbon intensity by 18% by 2025, which requires accurate forecasting of its emissions and their drivers. However, existing gray models have limitations in dealing with fluctuating data or long-time series data, and they often suffer from overfitting and poor generalization ability. Moreover, there is a lack of research and judgment on the future changes of emission drivers. To address these issues, this study proposes a fractional order gray adaptive rolling model (RFAGM(1,1)) that optimizes the background generation and incorporates the rolling mechanism. We apply RFAGM(1,1) to forecast China’s emissions, GDP, population, and consumption of raw coal, crude oil, and natural gas from 2020 to 2025. Our results show that RFAGM(1,1) achieves significantly higher accuracy than standard gray models, except for population. The projections indicate that China will meet the 18% carbon intensity reduction target by 2025. Furthermore, the LMDI decomposition reveals that economic growth and population changes have positive cumulative impacts on emission growth (245.68% and 11.95%, respectively), while energy intensity and structural changes have negative cumulative impacts (−151.60% and −6.02%, respectively). The improved forecast enables the evaluation of climate policies, while the factor analysis provides valuable insights for developing evidence-based strategies to achieve China’s carbon peaking and neutrality goals by 2030/2060.