[Analysis of the Influencing Factors of China's Carbon Emissions and Simulation of Peak Scenarios Based on Machine Learning].
摘要
In the context of the "dual carbon" goals， estimating carbon emissions and simulating peak scenarios has significant importance. However， accurately identifying the factors influencing carbon emissions and scientifically predicting the peak carbon emission time still present considerable challenges. This study focused on 30 provinces in China （data were not available for Hong Kong， Macau， Taiwan， and Tibet， China） and utilized the IPCC carbon emission coefficient method to estimate the total carbon emissions from 2000 to 2020， while also depicting their spatial and temporal distribution characteristics. Furthermore， various machine learning algorithms were employed to construct and select the optimal carbon emission estimation model， quantifying the main factors influencing carbon emissions in China. Finally， considering different development trends across the country and regions， scenario analysis and the carbon emission estimation model were used to predict changes in carbon emissions nationwide and in various regions over the next 15 years. The results follow： ① From 2000 to 2020， China's carbon emissions exhibited a "rapid then slow" growth pattern through time， with a spatial distribution characterized as "higher in the east and lower in the west， higher in the north and lower in the south." ② The main factors affecting China's carbon emissions include total energy consumption， energy structure， resident population， urbanization level， and the advancing industrial structure. ③ Under baseline， low-carbon， and high-carbon scenarios， carbon emissions nationwide and in various regions are projected to show an "increase then decrease" trend from 2021 to 2035. Notably， except for the northwest region under the high-carbon scenario， all regions are expected to achieve their carbon peak targets by 2030 under the three scenarios. This study provides valuable insights for formulating and implementing comprehensive control strategies for regional carbon peak emissions.