西北师范大学学报（自然科学版）2025，Vol.61Issue(5)：89-99,11.DOI:10.16783/j.cnki.nwnuz.2025.05.010
基于STIRPAT模型的甘肃省碳排放预测与影响因素研究
Research on carbon emission prediction and influencing factors in Gansu Province based on the STIRPAT model
摘要
Abstract
Based on the IPCC emission factor method,this study calculates the carbon emissions in Gansu Province from 1990 to 2021.It employs the LMDI decomposition method and the STIRPAT model to analyze the driving factors of carbon emissions,and uses a random forest algorithm to identify the key influencing factors.The PSO-BP neural network model is then applied to predict future trends under baseline,high-carbon,and low-carbon scenarios.The research findings indicate that energy utilization and the primary and secondary industries within the industrial structure have a restraining effect on carbon emissions.In contrast,factors such as the proportion of urban population,regional GDP,and energy consumption intensity significantly contribute to increased carbon emissions,with natural gas consumption being the largest contributor to carbon emission growth.Random forest analysis confirms that regional GDP,energy consumption intensity,urbanization level,industrial structure,and energy structure are the primary influencing factors.The introduction of the particle swarm optimization(PSO)algorithm effectively enhances the prediction accuracy of the BP neural network.The results show that under the baseline,high-carbon,and low-carbon scenarios,Gansu Province is expected to reach its carbon peak in 2025(250.4439 million tons),2030(260.9157 million tons),and 2035(320.5994 million tons),respectively.关键词
STIRPAT模型/碳排放/岭回归/随机森林/PSO-BP神经网络Key words
STIRPAT model/carbon emission/ridge regression/random forest/PSO-BP neural network分类
资源环境引用本文复制引用
周玉兰,陈明阳,谭婷婷..基于STIRPAT模型的甘肃省碳排放预测与影响因素研究[J].西北师范大学学报（自然科学版）,2025,61(5):89-99,11.基金项目
国家自然科学基金资助项目(12261080) （12261080）
